wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Merge branch 'main' into krisp-viva-vad-support

Browse Source
This commit is contained in:
Garegin Harutyunyan
2026-03-23 18:35:58 +04:00
committed by GitHub
parent c32240e14b ec3bd8c5b1
commit f1f51de962
741 changed files with 70990 additions and 21694 deletions
Download Patch File Download Diff File Expand all files Collapse all files

13
examples/foundational/01-say-one-thing-piper.py
View File

@@ -16,7 +16,7 @@ from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineTask
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.piper.tts import PiperTTSService
from pipecat.services.piper.tts import PiperHttpTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
@@ -24,9 +24,8 @@ from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_out_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_out_enabled=True),
@@ -39,8 +38,10 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# Create an HTTP session
async with aiohttp.ClientSession() as session:
tts = PiperTTSService(
base_url=os.getenv("PIPER_BASE_URL"), aiohttp_session=session, sample_rate=24000
tts = PiperHttpTTSService(
base_url=os.getenv("PIPER_BASE_URL"),
aiohttp_session=session,
sample_rate=24000,
)
task = PipelineTask(

9
examples/foundational/01-say-one-thing-rime.py
View File

@@ -23,9 +23,8 @@ from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_out_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_out_enabled=True),
@@ -40,8 +39,10 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async with aiohttp.ClientSession() as session:
tts = RimeHttpTTSService(
api_key=os.getenv("RIME_API_KEY", ""),
voice_id="rex",
aiohttp_session=session,
settings=RimeHttpTTSService.Settings(
voice="rex",
),
)
task = PipelineTask(

9
examples/foundational/01-say-one-thing.py
View File

@@ -23,9 +23,8 @@ from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_out_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_out_enabled=True),
@@ -38,7 +37,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
task = PipelineTask(

4
examples/foundational/01a-local-audio.py
View File

@@ -29,7 +29,9 @@ async def main():
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
pipeline = Pipeline([tts, transport.output()])

4
examples/foundational/01b-livekit-audio.py
View File

@@ -37,7 +37,9 @@ async def main():
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
runner = PipelineRunner()

5
examples/foundational/01c-nvidia-riva-tts.py
View File

@@ -23,9 +23,8 @@ from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_out_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_out_enabled=True),

27
examples/foundational/02-llm-say-one-thing.py
View File

@@ -25,9 +25,8 @@ from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_out_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_out_enabled=True),
@@ -40,17 +39,17 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are an LLM in a WebRTC session, and this is a 'hello world' demo. Say hello to the world.",
}
]
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
task = PipelineTask(
Pipeline([llm, tts, transport.output()]),
@@ -60,7 +59,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# Register an event handler so we can play the audio when the client joins
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
await task.queue_frames([LLMContextFrame(LLMContext(messages)), EndFrame()])
context = LLMContext()
context.add_message({"role": "user", "content": "Say hello to the world."})
await task.queue_frames([LLMContextFrame(context), EndFrame()])
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)

9
examples/foundational/03-still-frame.py
View File

@@ -23,9 +23,8 @@ from pipecat.transports.daily.transport import DailyParams
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
video_out_enabled=True,
@@ -46,7 +45,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# Create an HTTP session
async with aiohttp.ClientSession() as session:
imagegen = FalImageGenService(
params=FalImageGenService.InputParams(image_size="square_hd"),
settings=FalImageGenService.Settings(
image_size="square_hd",
),
aiohttp_session=session,
key=os.getenv("FAL_KEY"),
)

4
examples/foundational/03a-local-still-frame.py
View File

@@ -37,7 +37,9 @@ async def main():
)
imagegen = FalImageGenService(
params=FalImageGenService.InputParams(image_size="square_hd"),
settings=FalImageGenService.Settings(
image_size="square_hd",
),
aiohttp_session=session,
key=os.getenv("FAL_KEY"),
)

5
examples/foundational/03b-still-frame-imagen.py
View File

@@ -22,9 +22,8 @@ from pipecat.transports.daily.transport import DailyParams
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
video_out_enabled=True,

41
examples/foundational/04-transports-small-webrtc.py
View File

@@ -17,9 +17,7 @@ from fastapi.responses import RedirectResponse
from loguru import logger
from pipecat_ai_small_webrtc_prebuilt.frontend import SmallWebRTCPrebuiltUI
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -35,8 +33,6 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import TransportParams
from pipecat.transports.smallwebrtc.connection import IceServer, SmallWebRTCConnection
from pipecat.transports.smallwebrtc.transport import SmallWebRTCTransport
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -64,7 +60,6 @@ async def run_example(webrtc_connection: SmallWebRTCConnection):
params=TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
)
@@ -72,37 +67,33 @@ async def run_example(webrtc_connection: SmallWebRTCConnection):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -118,7 +109,7 @@ async def run_example(webrtc_connection: SmallWebRTCConnection):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

45
examples/foundational/04a-transports-daily.py
View File

@@ -12,9 +12,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -28,8 +26,6 @@ from pipecat.runner.daily import configure
from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.daily.transport import DailyParams, DailyTransport
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -49,44 +45,37 @@ async def main():
audio_in_enabled=True,
audio_out_enabled=True,
transcription_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
)
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"), model="gpt-4o")
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -102,7 +91,9 @@ async def main():
async def on_first_participant_joined(transport, participant):
await transport.capture_participant_transcription(participant["id"])
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_participant_left")

40
examples/foundational/04b-transports-livekit.py
View File

@@ -12,9 +12,7 @@ import sys
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import (
InterruptionFrame,
TranscriptionFrame,
@@ -35,8 +33,6 @@ from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.livekit.transport import LiveKitParams, LiveKitTransport
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -54,48 +50,40 @@ async def main():
params=LiveKitParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
)
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. "
"Your goal is to demonstrate your capabilities in a succinct way. "
"Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. "
"Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)

48
examples/foundational/05-sync-speech-and-image.py
View File

@@ -16,11 +16,12 @@ from pipecat.frames.frames import (
Frame,
LLMContextFrame,
LLMFullResponseStartFrame,
OutputImageRawFrame,
TextFrame,
)
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.sync_parallel_pipeline import SyncParallelPipeline
from pipecat.pipeline.sync_parallel_pipeline import FrameOrder, SyncParallelPipeline
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.sentence import SentenceAggregator
@@ -30,6 +31,7 @@ from pipecat.runner.utils import create_transport
from pipecat.services.cartesia.tts import CartesiaHttpTTSService
from pipecat.services.fal.image import FalImageGenService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.tts_service import TextAggregationMode
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
@@ -44,6 +46,18 @@ class MonthFrame(DataFrame):
return f"{self.name}(month: {self.month})"
class MarkImageForPlaybackSync(FrameProcessor):
"""Marks output image frames to be synchronized with audio playback."""
async def process_frame(self, frame: Frame, direction: FrameDirection):
await super().process_frame(frame, direction)
if isinstance(frame, OutputImageRawFrame):
frame.sync_with_audio = True
await self.push_frame(frame, direction)
class MonthPrepender(FrameProcessor):
def __init__(self):
super().__init__()
@@ -65,9 +79,8 @@ class MonthPrepender(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_out_enabled=True,
@@ -99,11 +112,19 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaHttpTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaHttpTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
# No need to aggregate by sentences (the default), as we already know we're getting full sentences
# (Otherwise the service will unnecessarily wait for follow-up input to confirm the sentence is complete,
# which, sadly, actually breaks the synchronization mechanism)
text_aggregation_mode=TextAggregationMode.TOKEN,
)
imagegen = FalImageGenService(
params=FalImageGenService.InputParams(image_size="square_hd"),
settings=FalImageGenService.Settings(
image_size="square_hd",
),
aiohttp_session=session,
key=os.getenv("FAL_KEY"),
)
@@ -116,17 +137,26 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# that, each pipeline runs concurrently and `SyncParallelPipeline` will
# wait for the input frame to be processed.
#
# We use `FrameOrder.PIPELINE` so that each synchronized batch of output
# frames is pushed in the order the pipelines are listed: image first,
# then audio. This ensures the transport receives the image before the
# audio frames it should accompany.
#
# Note that `SyncParallelPipeline` requires the last processor in each
# of the pipelines to be synchronous. In this case, we use
# `CartesiaHttpTTSService` and `FalImageGenService` which make HTTP
# `FalImageGenService` and `CartesiaHttpTTSService` which make HTTP
# requests and wait for the response.
pipeline = Pipeline(
[
llm, # LLM
sentence_aggregator, # Aggregates LLM output into full sentences
SyncParallelPipeline( # Run pipelines in parallel aggregating the result
[
imagegen, # Generate image
MarkImageForPlaybackSync(), # Mark image as needing sync w/audio during playback
],
[month_prepender, tts], # Create "Month: sentence" and output audio
[imagegen], # Generate image
frame_order=FrameOrder.PIPELINE,
),
transport.output(), # Transport output
]
@@ -149,7 +179,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
]:
messages = [
{
"role": "system",
"role": "user",
"content": f"Describe a nature photograph suitable for use in a calendar, for the month of {month}. Include only the image description with no preamble. Limit the description to one sentence, please.",
}
]

198
examples/foundational/05a-local-sync-speech-and-image.py
View File

@@ -1,198 +0,0 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import asyncio
import os
import sys
import tkinter as tk
import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.frames.frames import (
Frame,
LLMContextFrame,
OutputAudioRawFrame,
TextFrame,
TTSAudioRawFrame,
URLImageRawFrame,
)
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.sync_parallel_pipeline import SyncParallelPipeline
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.sentence import SentenceAggregator
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.services.cartesia.tts import CartesiaHttpTTSService
from pipecat.services.fal.image import FalImageGenService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.local.tk import TkLocalTransport, TkTransportParams
load_dotenv(override=True)
logger.remove(0)
logger.add(sys.stderr, level="DEBUG")
async def main():
async with aiohttp.ClientSession() as session:
tk_root = tk.Tk()
tk_root.title("Calendar")
runner = PipelineRunner()
async def get_month_data(month):
messages = [
{
"role": "system",
"content": f"Describe a nature photograph suitable for use in a calendar, for the month of {month}. Include only the image description with no preamble. Limit the description to one sentence, please.",
}
]
class ImageDescription(FrameProcessor):
def __init__(self):
super().__init__()
self.text = ""
async def process_frame(self, frame: Frame, direction: FrameDirection):
await super().process_frame(frame, direction)
if isinstance(frame, TextFrame):
self.text = frame.text
await self.push_frame(frame, direction)
class AudioGrabber(FrameProcessor):
def __init__(self):
super().__init__()
self.audio = bytearray()
self.frame = None
async def process_frame(self, frame: Frame, direction: FrameDirection):
await super().process_frame(frame, direction)
if isinstance(frame, TTSAudioRawFrame):
self.audio.extend(frame.audio)
self.frame = OutputAudioRawFrame(
bytes(self.audio), frame.sample_rate, frame.num_channels
)
await self.push_frame(frame, direction)
class ImageGrabber(FrameProcessor):
def __init__(self):
super().__init__()
self.frame = None
async def process_frame(self, frame: Frame, direction: FrameDirection):
await super().process_frame(frame, direction)
if isinstance(frame, URLImageRawFrame):
self.frame = frame
await self.push_frame(frame, direction)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
tts = CartesiaHttpTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
)
imagegen = FalImageGenService(
params=FalImageGenService.InputParams(image_size="square_hd"),
aiohttp_session=session,
key=os.getenv("FAL_KEY"),
)
sentence_aggregator = SentenceAggregator()
description = ImageDescription()
audio_grabber = AudioGrabber()
image_grabber = ImageGrabber()
# With `SyncParallelPipeline` we synchronize audio and images by
# pushing them basically in order (e.g. I1 A1 A1 A1 I2 A2 A2 A2 A2
# I3 A3). To do that, each pipeline runs concurrently and
# `SyncParallelPipeline` will wait for the input frame to be
# processed.
#
# Note that `SyncParallelPipeline` requires the last processor in
# each of the pipelines to be synchronous. In this case, we use
# `CartesiaHttpTTSService` and `FalImageGenService` which make HTTP
# requests and wait for the response.
pipeline = Pipeline(
[
llm, # LLM
sentence_aggregator, # Aggregates LLM output into full sentences
description, # Store sentence
SyncParallelPipeline(
[tts, audio_grabber], # Generate and store audio for the given sentence
[imagegen, image_grabber], # Generate and storeimage for the given sentence
),
]
)
task = PipelineTask(pipeline)
await task.queue_frame(LLMContextFrame(LLMContext(messages)))
await task.stop_when_done()
await runner.run(task)
return {
"month": month,
"text": description.text,
"image": image_grabber.frame,
"audio": audio_grabber.frame,
}
transport = TkLocalTransport(
tk_root,
TkTransportParams(
audio_out_enabled=True,
video_out_enabled=True,
video_out_width=1024,
video_out_height=1024,
),
)
pipeline = Pipeline([transport.output()])
task = PipelineTask(pipeline)
# We only specify a few months as we create tasks all at once and we
# might get rate limited otherwise.
months: list[str] = [
"January",
"February",
]
# We create one task per month. This will be executed concurrently.
month_tasks = [asyncio.create_task(get_month_data(month)) for month in months]
# Now we wait for each month task in the order they're completed. The
# benefit is we'll have as little delay as possible before the first
# month, and likely no delay between months, but the months won't
# display in order.
async def show_images(month_tasks):
for month_data_task in asyncio.as_completed(month_tasks):
data = await month_data_task
await task.queue_frames([data["image"], data["audio"]])
await runner.stop_when_done()
async def run_tk():
while not task.has_finished():
tk_root.update()
tk_root.update_idletasks()
await asyncio.sleep(0.1)
await asyncio.gather(runner.run(task), show_images(month_tasks), run_tk())
if __name__ == "__main__":
asyncio.run(main())

46
examples/foundational/06-listen-and-respond.py
View File

@@ -9,9 +9,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import Frame, LLMRunFrame, MetricsFrame
from pipecat.metrics.metrics import (
LLMUsageMetricsData,
@@ -36,8 +34,6 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -62,24 +58,20 @@ class MetricsLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -91,40 +83,36 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
ml = MetricsLogger()
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(),
stt,
context_aggregator.user(),
user_aggregator,
llm,
tts,
ml,
transport.output(),
context_aggregator.assistant(),
assistant_aggregator,
]
)
@@ -141,7 +129,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

45
examples/foundational/06a-image-sync.py
View File

@@ -10,9 +10,7 @@ from dotenv import load_dotenv
from loguru import logger
from PIL import Image
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import (
BotStartedSpeakingFrame,
BotStoppedSpeakingFrame,
@@ -36,8 +34,6 @@ from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -77,9 +73,8 @@ class ImageSyncAggregator(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
@@ -87,7 +82,6 @@ transport_params = {
video_out_enabled=True,
video_out_width=1024,
video_out_height=1024,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
@@ -95,7 +89,6 @@ transport_params = {
video_out_enabled=True,
video_out_width=1024,
video_out_height=1024,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -107,28 +100,24 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
image_sync_aggregator = ImageSyncAggregator(
os.path.join(os.path.dirname(__file__), "assets", "speaking.png"),
os.path.join(os.path.dirname(__file__), "assets", "waiting.png"),
@@ -138,12 +127,12 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
[
transport.input(),
stt,
context_aggregator.user(),
user_aggregator,
llm,
tts,
image_sync_aggregator,
transport.output(),
context_aggregator.assistant(),
assistant_aggregator,
]
)

121
examples/foundational/07-interruptible-cartesia-http.py
View File

@@ -6,12 +6,11 @@
import os
import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -29,30 +28,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -60,68 +53,68 @@ transport_params = {
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = CartesiaSTTService(api_key=os.getenv("CARTESIA_API_KEY"))
async with aiohttp.ClientSession() as session:
stt = CartesiaSTTService(api_key=os.getenv("CARTESIA_API_KEY"))
tts = CartesiaHttpTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
tts = CartesiaHttpTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
aiohttp_session=session,
settings=CartesiaHttpTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
),
)
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
]
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
await runner.run(task)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):

125
examples/foundational/07-interruptible-openai-responses.py Normal file
View File

@@ -0,0 +1,125 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.responses.llm import OpenAIResponsesLLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAIResponsesLLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAIResponsesLLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
context.add_message(
{"role": "developer", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

48
examples/foundational/07-interruptible.py
View File

@@ -9,9 +9,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -29,29 +27,23 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -63,37 +55,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -110,7 +98,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

43
examples/foundational/07a-interruptible-speechmatics-vad.py
View File

@@ -21,7 +21,6 @@ from pipecat.processors.aggregators.llm_response_universal import (
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.openai.base_llm import BaseOpenAILLMService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.speechmatics.stt import SpeechmaticsSTTService
from pipecat.services.speechmatics.tts import SpeechmaticsTTSService
@@ -33,9 +32,8 @@ from pipecat.turns.user_turn_strategies import ExternalUserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
@@ -94,7 +92,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async with aiohttp.ClientSession() as session:
stt = SpeechmaticsSTTService(
api_key=os.getenv("SPEECHMATICS_API_KEY"),
params=SpeechmaticsSTTService.InputParams(
settings=SpeechmaticsSTTService.Settings(
language=Language.EN,
turn_detection_mode=SpeechmaticsSTTService.TurnDetectionMode.ADAPTIVE,
# focus_speakers=["S1"],
@@ -105,33 +103,22 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = SpeechmaticsTTSService(
api_key=os.getenv("SPEECHMATICS_API_KEY"),
voice_id="sarah",
settings=SpeechmaticsTTSService.Settings(
voice="sarah",
),
aiohttp_session=session,
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
params=BaseOpenAILLMService.InputParams(temperature=0.75),
settings=OpenAILLMService.Settings(
temperature=0.75,
system_instruction="You are a helpful British assistant called Sarah in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Always include punctuation in your responses. Give very short replies - do not give longer replies unless strictly necessary. Respond to what the user said in a concise, funny, creative and helpful way. Use `<Sn/>` tags to identify different speakers - do not use tags in your replies. Do not respond to speakers within `<PASSIVE/>` tags unless explicitly asked to.",
),
)
messages = [
{
"role": "system",
"content": (
"You are a helpful British assistant called Sarah. "
"Your goal is to demonstrate your capabilities in a succinct way. "
"Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. "
"Always include punctuation in your responses. "
"Give very short replies - do not give longer replies unless strictly necessary. "
"Respond to what the user said in a concise, funny, creative and helpful way. "
"Use `<Sn/>` tags to identify different speakers - do not use tags in your replies. "
"Do not respond to speakers within `<PASSIVE/>` tags unless explicitly asked to. "
),
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(user_turn_strategies=ExternalUserTurnStrategies()),
)
@@ -140,11 +127,11 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -161,7 +148,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Say a short hello to the user."})
context.add_message({"role": "user", "content": "Say a short hello to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

57
examples/foundational/07a-interruptible-speechmatics.py
View File

@@ -10,9 +10,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -24,7 +22,6 @@ from pipecat.processors.aggregators.llm_response_universal import (
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.openai.base_llm import BaseOpenAILLMService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.speechmatics.stt import SpeechmaticsSTTService
from pipecat.services.speechmatics.tts import SpeechmaticsTTSService
@@ -32,29 +29,23 @@ from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -84,7 +75,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async with aiohttp.ClientSession() as session:
stt = SpeechmaticsSTTService(
api_key=os.getenv("SPEECHMATICS_API_KEY"),
params=SpeechmaticsSTTService.InputParams(
settings=SpeechmaticsSTTService.Settings(
language=Language.EN,
speaker_active_format="<{speaker_id}>{text}</{speaker_id}>",
),
@@ -92,51 +83,35 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = SpeechmaticsTTSService(
api_key=os.getenv("SPEECHMATICS_API_KEY"),
voice_id="sarah",
settings=SpeechmaticsTTSService.Settings(
voice="sarah",
),
aiohttp_session=session,
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
params=BaseOpenAILLMService.InputParams(temperature=0.75),
settings=OpenAILLMService.Settings(
temperature=0.75,
system_instruction="You are a helpful British assistant called Sarah in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Always include punctuation in your responses. Give very short replies - do not give longer replies unless strictly necessary. Respond to what the user said in a concise, funny, creative and helpful way. Use `<Sn/>` tags to identify different speakers - do not use tags in your replies. Do not respond to speakers within `<PASSIVE/>` tags unless explicitly asked to.",
),
)
messages = [
{
"role": "system",
"content": (
"You are a helpful British assistant called Sarah. "
"Your goal is to demonstrate your capabilities in a succinct way. "
"Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. "
"Always include punctuation in your responses. "
"Give very short replies - do not give longer replies unless strictly necessary. "
"Respond to what the user said in a concise, funny, creative and helpful way. "
"Use `<Sn/>` tags to identify different speakers - do not use tags in your replies."
),
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -153,7 +128,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Say a short hello to the user."})
context.add_message({"role": "user", "content": "Say a short hello to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

31
examples/foundational/07b-interruptible-langchain.py
View File

@@ -15,9 +15,7 @@ from langchain_core.runnables.history import RunnableWithMessageHistory
from langchain_openai import ChatOpenAI
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMMessagesUpdateFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -35,8 +33,6 @@ from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -50,24 +46,20 @@ def get_session_history(session_id: str) -> BaseChatMessageHistory:
return message_store[session_id]
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -79,15 +71,16 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
prompt = ChatPromptTemplate.from_messages(
[
(
"system",
"Be nice and helpful. Answer very briefly and without special characters like `#` or `*`. "
"Your response will be synthesized to voice and those characters will create unnatural sounds.",
"You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
MessagesPlaceholder("chat_history"),
("human", "{input}"),
@@ -103,24 +96,20 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
lc = LangchainProcessor(history_chain)
context = LLMContext()
context_aggregator = LLMContextAggregatorPair(
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
lc, # Langchain
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)

46
examples/foundational/07c-interruptible-deepgram-flux.py
View File

@@ -10,6 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -32,9 +33,8 @@ from pipecat.turns.user_turn_strategies import ExternalUserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
@@ -56,35 +56,43 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramFluxSTTService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
params=DeepgramFluxSTTService.InputParams(min_confidence=0.3),
settings=DeepgramFluxSTTService.Settings(
min_confidence=0.3,
),
)
tts = DeepgramTTSService(api_key=os.getenv("DEEPGRAM_API_KEY"), voice="aura-2-andromeda-en")
tts = DeepgramTTSService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
settings=DeepgramTTSService.Settings(
voice="aura-2-andromeda-en",
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(user_turn_strategies=ExternalUserTurnStrategies()),
user_params=LLMUserAggregatorParams(
user_turn_strategies=ExternalUserTurnStrategies(),
vad_analyzer=SileroVADAnalyzer(),
),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -101,7 +109,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

52
examples/foundational/07c-interruptible-deepgram-http.py
View File

@@ -11,9 +11,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,30 +29,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -67,40 +59,34 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = DeepgramHttpTTSService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
voice="aura-2-andromeda-en",
settings=DeepgramHttpTTSService.Settings(
voice="aura-2-andromeda-en",
),
aiohttp_session=session,
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -117,7 +103,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

62
examples/foundational/07c-interruptible-deepgram-sagemaker.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -24,36 +22,30 @@ from pipecat.processors.aggregators.llm_response_universal import (
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.aws.llm import AWSBedrockLLMService
from pipecat.services.deepgram.stt_sagemaker import DeepgramSageMakerSTTService
from pipecat.services.deepgram.tts import DeepgramTTSService
from pipecat.services.aws.llm import AWSBedrockLLMService, AWSBedrockLLMSettings
from pipecat.services.deepgram.sagemaker.stt import DeepgramSageMakerSTTService
from pipecat.services.deepgram.sagemaker.tts import DeepgramSageMakerTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -66,44 +58,46 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# - AWS credentials configured (via environment variables or AWS CLI)
# - A deployed SageMaker endpoint with Deepgram model
stt = DeepgramSageMakerSTTService(
endpoint_name=os.getenv("SAGEMAKER_ENDPOINT_NAME"),
endpoint_name=os.getenv("SAGEMAKER_STT_ENDPOINT_NAME"),
region=os.getenv("AWS_REGION"),
)
tts = DeepgramTTSService(api_key=os.getenv("DEEPGRAM_API_KEY"), voice="aura-2-andromeda-en")
# Initialize Deepgram SageMaker TTS Service
# This requires:
# - AWS credentials configured (via environment variables or AWS CLI)
# - A deployed SageMaker endpoint with Deepgram TTS model
tts = DeepgramSageMakerTTSService(
endpoint_name=os.getenv("SAGEMAKER_TTS_ENDPOINT_NAME"),
region=os.getenv("AWS_REGION"),
settings=DeepgramSageMakerTTSService.Settings(
voice="aura-2-andromeda-en",
),
)
llm = AWSBedrockLLMService(
aws_region=os.getenv("AWS_REGION"),
model="us.amazon.nova-pro-v1:0",
params=AWSBedrockLLMService.InputParams(temperature=0.8),
settings=AWSBedrockLLMSettings(
model="us.amazon.nova-pro-v1:0",
temperature=0.8,
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -120,7 +114,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

42
examples/foundational/07c-interruptible-deepgram-vad.py
View File

@@ -7,7 +7,6 @@
import os
from deepgram import LiveOptions
from dotenv import load_dotenv
from loguru import logger
@@ -33,9 +32,8 @@ from pipecat.turns.user_turn_strategies import ExternalUserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
@@ -57,22 +55,28 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
live_options=LiveOptions(vad_events=True, utterance_end_ms="1000"),
settings=DeepgramSTTService.Settings(
vad_events=True,
utterance_end_ms="1000",
),
)
tts = DeepgramTTSService(api_key=os.getenv("DEEPGRAM_API_KEY"), voice="aura-2-andromeda-en")
tts = DeepgramTTSService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
settings=DeepgramTTSService.Settings(
voice="aura-2-andromeda-en",
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(user_turn_strategies=ExternalUserTurnStrategies()),
)
@@ -81,11 +85,11 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -102,7 +106,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

53
examples/foundational/07c-interruptible-deepgram.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,30 +28,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -63,36 +55,35 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = DeepgramTTSService(api_key=os.getenv("DEEPGRAM_API_KEY"), voice="aura-2-andromeda-en")
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
tts = DeepgramTTSService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
settings=DeepgramTTSService.Settings(
voice="aura-2-andromeda-en",
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -109,7 +100,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

52
examples/foundational/07d-interruptible-elevenlabs-http.py
View File

@@ -11,9 +11,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,30 +29,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -71,40 +63,34 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = ElevenLabsHttpTTSService(
api_key=os.getenv("ELEVENLABS_API_KEY", ""),
voice_id=os.getenv("ELEVENLABS_VOICE_ID", ""),
aiohttp_session=session,
settings=ElevenLabsHttpTTSService.Settings(
voice=os.getenv("ELEVENLABS_VOICE_ID", ""),
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -121,7 +107,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

48
examples/foundational/07d-interruptible-elevenlabs.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,30 +28,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -65,37 +57,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = ElevenLabsTTSService(
api_key=os.getenv("ELEVENLABS_API_KEY", ""),
voice_id=os.getenv("ELEVENLABS_VOICE_ID", ""),
settings=ElevenLabsTTSService.Settings(
voice=os.getenv("ELEVENLABS_VOICE_ID", ""),
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -112,7 +100,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

40
examples/foundational/07f-interruptible-azure-http.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,29 +28,23 @@ from pipecat.services.azure.tts import AzureHttpTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -73,35 +65,27 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
llm = AzureLLMService(
api_key=os.getenv("AZURE_CHATGPT_API_KEY"),
endpoint=os.getenv("AZURE_CHATGPT_ENDPOINT"),
model=os.getenv("AZURE_CHATGPT_MODEL"),
settings=AzureLLMService.Settings(
model=os.getenv("AZURE_CHATGPT_MODEL"),
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -118,7 +102,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

40
examples/foundational/07f-interruptible-azure.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,29 +28,23 @@ from pipecat.services.azure.tts import AzureTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -73,35 +65,27 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
llm = AzureLLMService(
api_key=os.getenv("AZURE_CHATGPT_API_KEY"),
endpoint=os.getenv("AZURE_CHATGPT_ENDPOINT"),
model=os.getenv("AZURE_CHATGPT_MODEL"),
settings=AzureLLMService.Settings(
model=os.getenv("AZURE_CHATGPT_MODEL"),
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -118,7 +102,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

131
examples/foundational/07g-interruptible-openai-http.py Normal file
View File

@@ -0,0 +1,131 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.openai.stt import OpenAISTTService
from pipecat.services.openai.tts import OpenAITTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = OpenAISTTService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAISTTService.Settings(
model="gpt-4o-transcribe",
prompt="Expect words related to dogs, such as breed names.",
),
)
tts = OpenAITTSService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAITTSService.Settings(
voice="ballad",
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are very knowledgable about dogs. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
audio_out_sample_rate=24000,
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

65
examples/foundational/07g-interruptible-openai.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -25,34 +23,29 @@ from pipecat.processors.aggregators.llm_response_universal import (
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.openai.stt import OpenAISTTService
from pipecat.services.openai.stt import OpenAIRealtimeSTTService
from pipecat.services.openai.tts import OpenAITTSService
from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -60,42 +53,44 @@ transport_params = {
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = OpenAISTTService(
stt = OpenAIRealtimeSTTService(
api_key=os.getenv("OPENAI_API_KEY"),
model="gpt-4o-transcribe",
prompt="Expect words related to dogs, such as breed names.",
settings=OpenAIRealtimeSTTService.Settings(
model="gpt-4o-transcribe",
prompt="Expect words related to dogs, such as breed names.",
language=Language.EN,
),
)
tts = OpenAITTSService(api_key=os.getenv("OPENAI_API_KEY"), voice="ballad")
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are very knowledgable about dogs. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
tts = OpenAITTSService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAITTSService.Settings(
voice="ballad",
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are very knowledgable about dogs. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -113,7 +108,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

42
examples/foundational/07h-interruptible-openpipe.py
View File

@@ -11,9 +11,7 @@ import time
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,29 +29,23 @@ from pipecat.services.openpipe.llm import OpenPipeLLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -65,7 +57,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
timestamp = int(time.time())
@@ -73,34 +67,26 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
api_key=os.getenv("OPENAI_API_KEY"),
openpipe_api_key=os.getenv("OPENPIPE_API_KEY"),
tags={"conversation_id": f"pipecat-{timestamp}"},
settings=OpenPipeLLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -117,7 +103,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

52
examples/foundational/07i-interruptible-xtts.py
View File

@@ -11,9 +11,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,29 +29,23 @@ from pipecat.services.xtts.tts import XTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -67,40 +59,34 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = XTTSService(
aiohttp_session=session,
voice_id="Claribel Dervla",
settings=XTTSService.Settings(
voice="Claribel Dervla",
),
base_url="http://localhost:8000",
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -117,7 +103,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

138
examples/foundational/07j-interruptible-gladia-vad.py Normal file
View File

@@ -0,0 +1,138 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.gladia.config import LanguageConfig
from pipecat.services.gladia.stt import GladiaSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_turn_strategies import ExternalUserTurnStrategies
load_dotenv(override=True)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = GladiaSTTService(
api_key=os.getenv("GLADIA_API_KEY", ""),
region=os.getenv("GLADIA_REGION"),
settings=GladiaSTTService.Settings(
language_config=LanguageConfig(
languages=[Language.EN],
),
enable_vad=True,
),
)
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY", ""),
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY", ""),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=ExternalUserTurnStrategies(),
vad_analyzer=SileroVADAnalyzer(),
),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

52
examples/foundational/07j-interruptible-gladia.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -25,36 +23,30 @@ from pipecat.processors.aggregators.llm_response_universal import (
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.gladia.config import GladiaInputParams, LanguageConfig
from pipecat.services.gladia.config import LanguageConfig
from pipecat.services.gladia.stt import GladiaSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -65,7 +57,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = GladiaSTTService(
api_key=os.getenv("GLADIA_API_KEY", ""),
region=os.getenv("GLADIA_REGION"),
params=GladiaInputParams(
settings=GladiaSTTService.Settings(
language_config=LanguageConfig(
languages=[Language.EN],
)
@@ -74,37 +66,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY", ""),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY", ""))
messages = [
{
"role": "system",
"content": f"You are a helpful LLM. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY", ""),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -121,7 +109,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

54
examples/foundational/07k-interruptible-lmnt.py
View File

@@ -10,10 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.base_smart_turn import SmartTurnParams
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,29 +28,23 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -63,36 +54,35 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = LmntTTSService(api_key=os.getenv("LMNT_API_KEY"), voice_id="morgan")
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
tts = LmntTTSService(
api_key=os.getenv("LMNT_API_KEY"),
settings=LmntTTSService.Settings(
voice="morgan",
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User respones
user_aggregator, # User respones
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -109,7 +99,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

41
examples/foundational/07l-interruptible-groq.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,29 +28,23 @@ from pipecat.services.groq.tts import GroqTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -63,37 +55,30 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = GroqSTTService(api_key=os.getenv("GROQ_API_KEY"))
llm = GroqLLMService(
api_key=os.getenv("GROQ_API_KEY"), model="meta-llama/llama-4-maverick-17b-128e-instruct"
api_key=os.getenv("GROQ_API_KEY"),
settings=GroqLLMService.Settings(
model="llama-3.1-8b-instant",
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
tts = GroqTTSService(api_key=os.getenv("GROQ_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -110,7 +95,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

34
examples/foundational/07m-interruptible-aws-strands.py
View File

@@ -8,7 +8,6 @@
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMMessagesAppendFrame
from pipecat.pipeline.pipeline import Pipeline
@@ -27,8 +26,6 @@ from pipecat.services.aws.tts import AWSPollyTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
# Strands agent setup
try:
@@ -41,24 +38,20 @@ except ImportError:
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
}
@@ -102,13 +95,16 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = AWSPollyTTSService(
region="us-west-2", # only specific regions support generative TTS
voice_id="Joanna",
params=AWSPollyTTSService.InputParams(engine="generative", rate="1.1"),
settings=AWSPollyTTSService.Settings(
voice="Joanna",
engine="generative",
rate="1.1",
),
)
# Create Strands agent processor
try:
agent = build_agent(model_id="us.anthropic.claude-3-5-haiku-20241022-v1:0", max_tokens=8000)
agent = build_agent(model_id="us.anthropic.claude-sonnet-4-6", max_tokens=8000)
llm = StrandsAgentsProcessor(agent=agent)
logger.info("Successfully created Strands agent for NAB customer service coaching")
except Exception as e:
@@ -120,24 +116,20 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
# Setup context aggregators for message handling
context = LLMContext()
context_aggregator = LLMContextAggregatorPair(
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # Speech-to-text
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # Strands Agents processor
tts, # Text-to-speech
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -160,7 +152,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
messages=[
{
"role": "user",
"content": f"Greet the user and introduce yourself.",
"content": f"Greet the user and introduce yourself. Don't use emojis.",
}
],
run_llm=True,

49
examples/foundational/07m-interruptible-aws.py
View File

@@ -8,9 +8,7 @@
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -28,29 +26,23 @@ from pipecat.services.aws.tts import AWSPollyTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -62,42 +54,37 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = AWSPollyTTSService(
region="us-west-2", # only specific regions support generative TTS
voice_id="Joanna",
params=AWSPollyTTSService.InputParams(engine="generative", rate="1.1"),
settings=AWSPollyTTSService.Settings(
voice="Joanna",
engine="generative",
rate="1.1",
),
)
llm = AWSBedrockLLMService(
aws_region="us-west-2",
model="us.anthropic.claude-haiku-4-5-20251001-v1:0",
params=AWSBedrockLLMService.InputParams(temperature=0.8),
settings=AWSBedrockLLMService.Settings(
model="us.anthropic.claude-sonnet-4-6",
temperature=0.8,
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -114,7 +101,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "user", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

52
examples/foundational/07n-interruptible-gemini-image.py
View File

@@ -25,9 +25,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -45,15 +43,11 @@ from pipecat.services.google.tts import GoogleTTSService
from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
@@ -61,7 +55,6 @@ transport_params = {
video_out_enabled=True,
video_out_width=1024,
video_out_height=1024,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
@@ -69,7 +62,6 @@ transport_params = {
video_out_enabled=True,
video_out_width=1024,
video_out_height=1024,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -78,48 +70,44 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = GoogleSTTService(
params=GoogleSTTService.InputParams(languages=Language.EN_US),
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
settings=GoogleSTTService.Settings(
languages=[Language.EN_US],
),
)
tts = GoogleTTSService(
voice_id="en-US-Chirp3-HD-Charon",
params=GoogleTTSService.InputParams(language=Language.EN_US),
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
settings=GoogleTTSService.Settings(
voice="en-US-Chirp3-HD-Charon",
language=Language.EN_US,
),
)
llm = GoogleLLMService(
api_key=os.getenv("GOOGLE_API_KEY"),
model="gemini-2.5-flash-image",
# model="gemini-3-pro-image-preview", # A more powerful model, but slower
settings=GoogleLLMService.Settings(
model="gemini-2.5-flash-image",
# model="gemini-3-pro-image-preview", # A more powerful model, but slower,
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # Gemini TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -136,7 +124,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation with a styled introduction
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

57
examples/foundational/07n-interruptible-gemini.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,29 +29,23 @@ from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -62,15 +54,17 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot with Gemini TTS")
stt = GoogleSTTService(
params=GoogleSTTService.InputParams(languages=Language.EN_US),
settings=GoogleSTTService.Settings(
languages=[Language.EN_US],
),
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
)
tts = GeminiTTSService(
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
model="gemini-2.5-flash-tts",
voice_id="Charon",
params=GeminiTTSService.InputParams(
settings=GeminiTTSService.Settings(
model="gemini-2.5-flash-tts",
voice="Charon",
language=Language.EN_US,
prompt="You are a helpful AI assistant. Speak in a natural, conversational tone.",
),
@@ -79,13 +73,8 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
llm = GoogleLLMService(
api_key=os.getenv("GOOGLE_API_KEY"),
model="gemini-2.5-flash",
)
# System message that instructs the AI on how to speak
messages = [
{
"role": "system",
"content": """You are a helpful AI assistant in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way.
settings=GoogleLLMService.Settings(
system_instruction="""You are a helpful assistant in a voice conversation.
IMPORTANT: You're using Gemini TTS which supports expressive markup tags. You can use these tags in your responses:
- [sigh] - Insert a sigh sound
@@ -102,29 +91,25 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
- "[whispering] Let me tell you a secret."
- "The answer is... [long pause] ...42!"
Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.""",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Keep responses concise. Respond to what the user said in a creative and helpful way.""",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # Gemini TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -141,9 +126,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation
messages.append(
context.add_message(
{
"role": "system",
"role": "user",
"content": "You are an AI assistant. You can help with a variety of tasks. Introduce yourself and ask the user what they would like to know.",
}
)

58
examples/foundational/07n-interruptible-google-http.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,29 +29,23 @@ from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -62,52 +54,48 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = GoogleSTTService(
params=GoogleSTTService.InputParams(languages=Language.EN_US, model="chirp_3"),
settings=GoogleSTTService.Settings(
languages=[Language.EN_US],
# Add model to use a specific model
# model="chirp_3",
),
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
location="us",
)
tts = GoogleHttpTTSService(
voice_id="en-US-Chirp3-HD-Charon",
params=GoogleHttpTTSService.InputParams(language=Language.EN_US),
settings=GoogleHttpTTSService.Settings(
voice="en-US-Chirp3-HD-Charon",
language=Language.EN_US,
),
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
)
llm = GoogleLLMService(
api_key=os.getenv("GOOGLE_API_KEY"),
model="gemini-2.5-flash",
# force a certain amount of thinking if you want it
# params=GoogleLLMService.InputParams(
# thinking=GoogleLLMService.ThinkingConfig(thinking_budget=4096)
# ),
settings=GoogleLLMService.Settings(
model="gemini-2.5-flash",
# force a certain amount of thinking if you want it
# thinking=GoogleLLMService.ThinkingConfig(thinking_budget=4096)
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User respones
user_aggregator, # User respones
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -124,7 +112,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

58
examples/foundational/07n-interruptible-google.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,29 +29,23 @@ from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -62,52 +54,48 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = GoogleSTTService(
params=GoogleSTTService.InputParams(languages=Language.EN_US, model="chirp_3"),
settings=GoogleSTTService.Settings(
languages=[Language.EN_US],
# Add model to use a specific model
# model="chirp_3",
),
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
location="us",
)
tts = GoogleTTSService(
voice_id="en-US-Chirp3-HD-Charon",
params=GoogleTTSService.InputParams(language=Language.EN_US),
settings=GoogleTTSService.Settings(
voice="en-US-Chirp3-HD-Charon",
language=Language.EN_US,
),
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
)
llm = GoogleLLMService(
api_key=os.getenv("GOOGLE_API_KEY"),
model="gemini-2.5-flash",
# force a certain amount of thinking if you want it
# params=GoogleLLMService.InputParams(
# thinking=GoogleLLMService.ThinkingConfig(thinking_budget=4096)
# ),
settings=GoogleLLMService.Settings(
model="gemini-2.5-flash",
# force a certain amount of thinking if you want it
# thinking=GoogleLLMService.ThinkingConfig(thinking_budget=4096),
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User respones
user_aggregator, # User respones
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -124,7 +112,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

178
examples/foundational/07o-interruptible-assemblyai-turn-detection.py Normal file
View File

@@ -0,0 +1,178 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.assemblyai.stt import AssemblyAISTTService
from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_turn_strategies import ExternalUserTurnStrategies
load_dotenv(override=True)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
"""AssemblyAI u3-rt-pro with Built-in Turn Detection
This example demonstrates using AssemblyAI's u3-rt-pro Speech-to-Text model
with AssemblyAI's built-in turn detection for more natural conversation flow.
Key features:
1. AssemblyAI Turn Detection
- Set `vad_force_turn_endpoint=False` to use AssemblyAI's built-in turn detection
- AssemblyAI's model determines when user starts/stops speaking
- Uses `ExternalUserTurnStrategies` to delegate turn control to AssemblyAI
- More natural turn detection based on speech patterns and pauses
2. Advanced Turn Detection Tuning
- `min_turn_silence`: Minimum silence (ms) when confident about end-of-turn.
Lower values = faster responses. Default: 100ms
- `max_turn_silence`: Maximum silence (ms) before forcing end-of-turn.
Prevents long pauses. Default: 1000ms
3. Prompt-Based Transcription Enhancement
- Use `prompt` parameter to improve accuracy for specific names/terms
- Particularly useful for proper nouns, technical terms, domain vocabulary
- Example: "Names: Xiomara, Saoirse, Krzystof. Technical terms: API, OAuth."
4. Speaker Diarization (Optional)
- Enable with `speaker_labels=True`
- Automatically identifies different speakers in multi-party conversations
- TranscriptionFrame includes speaker_id field (e.g., "Speaker A", "Speaker B")
5. Language Detection (Optional, multilingual model only)
- Enable with `language_detection=True`
- Automatically detects spoken language
- Available with universal-streaming-multilingual model
For more information: https://www.assemblyai.com/docs/speech-to-text/streaming
"""
logger.info(f"Starting bot")
stt = AssemblyAISTTService(
api_key=os.getenv("ASSEMBLYAI_API_KEY"),
vad_force_turn_endpoint=False, # Use AssemblyAI's built-in turn detection
settings=AssemblyAISTTService.Settings(
model="u3-rt-pro",
# Optional: Tune turn detection timing (defaults shown below)
# min_turn_silence=100, # Default
# max_turn_silence=1000, # Default
# Optional: Boost accuracy for specific names/terms
# keyterms_prompt=["Xiomara", "Saoirse", "Krzystof", "API", "OAuth"],
# Optional: Enable speaker diarization
# speaker_labels=True,
),
)
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=ExternalUserTurnStrategies(),
vad_analyzer=SileroVADAnalyzer(),
),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

48
examples/foundational/07o-interruptible-assemblyai.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,30 +28,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -67,37 +59,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -114,7 +102,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

48
examples/foundational/07p-interruptible-krisp-viva.py
View File

@@ -28,10 +28,11 @@ from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.filters.krisp_viva_filter import KrispVivaFilter
from pipecat.audio.turn.krisp_viva_turn import KrispTurnParams, KrispVivaTurn
from pipecat.audio.turn.krisp_viva_turn import KrispVivaTurn
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.metrics.metrics import TurnMetricsData
from pipecat.observers.loggers.metrics_log_observer import MetricsLogObserver
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
@@ -42,8 +43,8 @@ from pipecat.processors.aggregators.llm_response_universal import (
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.deepgram.tts import DeepgramTTSService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
@@ -53,9 +54,11 @@ from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
krisp_viva_filter = KrispVivaFilter()
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
@@ -83,24 +86,28 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = DeepgramTTSService(api_key=os.getenv("DEEPGRAM_API_KEY"), voice="aura-helios-en")
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=KrispVivaTurn())]
),
vad_analyzer=SileroVADAnalyzer(),
),
)
@@ -108,11 +115,11 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -123,13 +130,14 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
observers=[MetricsLogObserver(include_metrics={TurnMetricsData})],
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

53
examples/foundational/07p-interruptible-krisp.py
View File

@@ -11,9 +11,7 @@ from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.filters.krisp_filter import KrispFilter
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,31 +29,25 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
audio_in_filter=KrispFilter(),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
audio_in_filter=KrispFilter(),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
audio_in_filter=KrispFilter(),
),
}
@@ -66,36 +58,35 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = DeepgramTTSService(api_key=os.getenv("DEEPGRAM_API_KEY"), voice="aura-helios-en")
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
tts = DeepgramTTSService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
settings=DeepgramTTSService.Settings(
voice="aura-helios-en",
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -112,7 +103,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

53
examples/foundational/07q-interruptible-rime-http.py
View File

@@ -11,9 +11,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,30 +29,24 @@ from pipecat.services.rime.tts import RimeHttpTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -68,41 +60,36 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = RimeHttpTTSService(
api_key=os.getenv("RIME_API_KEY", ""),
voice_id="luna",
settings=RimeHttpTTSService.Settings(
voice="luna",
model="arcana",
),
model="arcana",
aiohttp_session=session,
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -119,7 +106,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

48
examples/foundational/07q-interruptible-rime.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,29 +28,23 @@ from pipecat.services.rime.tts import RimeTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -64,37 +56,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = RimeTTSService(
api_key=os.getenv("RIME_API_KEY", ""),
voice_id="rex",
settings=RimeTTSService.Settings(
voice="luna",
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -111,7 +99,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

41
examples/foundational/07r-interruptible-nvidia.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,29 +28,23 @@ from pipecat.services.nvidia.tts import NvidiaTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -63,37 +55,30 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = NvidiaSTTService(api_key=os.getenv("NVIDIA_API_KEY"))
llm = NvidiaLLMService(
api_key=os.getenv("NVIDIA_API_KEY"), model="meta/llama-3.1-405b-instruct"
api_key=os.getenv("NVIDIA_API_KEY"),
settings=NvidiaLLMService.Settings(
model="meta/llama-3.3-70b-instruct",
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
tts = NvidiaTTSService(api_key=os.getenv("NVIDIA_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -110,7 +95,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

61
examples/foundational/07s-interruptible-google-audio-in.py
View File

@@ -12,9 +12,7 @@ from dotenv import load_dotenv
from google.genai.types import Content, Part
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import (
Frame,
InputAudioRawFrame,
@@ -44,15 +42,13 @@ from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
marker = "|----|"
system_message = f"""
You are a helpful LLM in a WebRTC call. Your goals are to be helpful and brief in your responses.
You are a helpful LLM in a voice call. Your goals are to be helpful and brief in your responses.
You are expert at transcribing audio to text. You will receive a mixture of audio and text input. When
asked to transcribe what the user said, output an exact, word-for-word transcription.
@@ -100,7 +96,7 @@ class UserAudioCollector(FrameProcessor):
self._user_speaking = True
elif isinstance(frame, UserStoppedSpeakingFrame):
self._user_speaking = False
self._context.add_audio_frames_message(audio_frames=self._audio_frames)
await self._context.add_audio_frames_message(audio_frames=self._audio_frames)
await self._user_context_aggregator.push_frame(LLMRunFrame())
elif isinstance(frame, InputAudioRawFrame):
@@ -197,24 +193,20 @@ class TranscriptionContextFixup(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -224,40 +216,29 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
llm = GoogleLLMService(
api_key=os.getenv("GOOGLE_API_KEY"),
model="gemini-2.5-flash",
# force a certain amount of thinking if you want it
# params=GoogleLLMService.InputParams(
# thinking=GoogleLLMService.ThinkingConfig(thinking_budget=4096)
# ),
settings=GoogleLLMService.Settings(
model="gemini-2.5-flash",
system_instruction=system_message,
# force a certain amount of thinking if you want it
# thinking=GoogleLLMService.ThinkingConfig(thinking_budget=4096)
),
)
tts = GoogleTTSService(
voice_id="en-US-Chirp3-HD-Charon",
settings=GoogleTTSService.Settings(
voice="en-US-Chirp3-HD-Charon",
language=Language.EN_US,
),
params=GoogleTTSService.InputParams(language=Language.EN_US),
credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"),
)
messages = [
{
"role": "system",
"content": system_message,
},
{
"role": "user",
"content": "Start by saying hello.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
audio_collector = UserAudioCollector(context, context_aggregator.user())
audio_collector = UserAudioCollector(context, user_aggregator)
pull_transcript_out_of_llm_output = TranscriptExtractor(context)
fixup_context_messages = TranscriptionContextFixup(context)
@@ -265,12 +246,12 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
[
transport.input(), # Transport user input
audio_collector,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
pull_transcript_out_of_llm_output,
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
fixup_context_messages,
]
)
@@ -288,7 +269,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

48
examples/foundational/07t-interruptible-fish.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,30 +28,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -65,37 +57,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = FishAudioTTSService(
api_key=os.getenv("FISH_API_KEY"),
model="4ce7e917cedd4bc2bb2e6ff3a46acaa1", # Barack Obama
settings=FishAudioTTSService.Settings(
voice="4ce7e917cedd4bc2bb2e6ff3a46acaa1", # Barack Obama
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -112,7 +100,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

52
examples/foundational/07v-interruptible-neuphonic-http.py
View File

@@ -11,9 +11,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,30 +29,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -68,40 +60,34 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = NeuphonicHttpTTSService(
api_key=os.getenv("NEUPHONIC_API_KEY"),
voice_id="fc854436-2dac-4d21-aa69-ae17b54e98eb", # Emily
settings=NeuphonicHttpTTSService.Settings(
voice="fc854436-2dac-4d21-aa69-ae17b54e98eb", # Emily
),
aiohttp_session=session,
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -118,7 +104,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

48
examples/foundational/07v-interruptible-neuphonic.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,29 +28,23 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -64,37 +56,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = NeuphonicTTSService(
api_key=os.getenv("NEUPHONIC_API_KEY"),
voice_id="fc854436-2dac-4d21-aa69-ae17b54e98eb", # Emily
settings=NeuphonicTTSService.Settings(
voice="fc854436-2dac-4d21-aa69-ae17b54e98eb", # Emily
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -111,7 +99,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

125
examples/foundational/07w-interruptible-fal.py
View File

@@ -7,12 +7,11 @@
import os
import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,30 +29,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -61,70 +54,70 @@ transport_params = {
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = FalSTTService(
api_key=os.getenv("FAL_KEY"),
)
async with aiohttp.ClientSession() as session:
stt = FalSTTService(
api_key=os.getenv("FAL_KEY"),
aiohttp_session=session,
)
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
),
)
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
]
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
await runner.run(task)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):

41
examples/foundational/07x-interruptible-local.py
View File

@@ -11,9 +11,7 @@ import sys
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -27,8 +25,6 @@ from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.local.audio import LocalAudioTransport, LocalAudioTransportParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -41,7 +37,6 @@ async def main():
LocalAudioTransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
)
)
@@ -49,37 +44,33 @@ async def main():
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -91,7 +82,7 @@ async def main():
),
)
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
runner = PipelineRunner()

52
examples/foundational/07y-interruptible-minimax.py
View File

@@ -11,9 +11,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -32,30 +30,24 @@ from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -71,39 +63,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
api_key=os.getenv("MINIMAX_API_KEY", ""),
group_id=os.getenv("MINIMAX_GROUP_ID", ""),
aiohttp_session=session,
params=MiniMaxHttpTTSService.InputParams(language=Language.EN),
settings=MiniMaxHttpTTSService.Settings(
language=Language.EN,
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -120,7 +106,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

56
examples/foundational/07z-interruptible-sarvam-http.py
View File

@@ -11,9 +11,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -32,30 +30,24 @@ from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -67,45 +59,41 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async with aiohttp.ClientSession() as session:
stt = SarvamSTTService(
api_key=os.getenv("SARVAM_API_KEY"),
model="saarika:v2.5",
settings=SarvamSTTService.Settings(
model="saarika:v2.5",
),
)
tts = SarvamHttpTTSService(
api_key=os.getenv("SARVAM_API_KEY"),
aiohttp_session=session,
params=SarvamHttpTTSService.InputParams(language=Language.EN),
settings=SarvamHttpTTSService.Settings(
language=Language.EN_IN,
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -122,7 +110,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

56
examples/foundational/07z-interruptible-sarvam.py
View File

@@ -9,9 +9,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -29,30 +27,24 @@ from pipecat.services.sarvam.tts import SarvamTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -62,42 +54,40 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = SarvamSTTService(
api_key=os.getenv("SARVAM_API_KEY"),
model="saarika:v2.5",
settings=SarvamSTTService.Settings(
model="saarika:v2.5",
),
)
tts = SarvamTTSService(
api_key=os.getenv("SARVAM_API_KEY"),
model="bulbul:v2",
voice_id="manisha",
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
settings=SarvamTTSService.Settings(
model="bulbul:v2",
voice="manisha",
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -113,7 +103,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
# Optionally, you can wait for 30 seconds and then change the voice.

50
examples/foundational/07aa-interruptible-soniox.py → examples/foundational/07za-interruptible-soniox.py
View File

@@ -10,9 +10,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -27,11 +25,10 @@ from pipecat.runner.utils import create_transport
from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.soniox.stt import SonioxSTTService
from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -39,17 +36,14 @@ transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -59,41 +53,43 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = SonioxSTTService(
api_key=os.getenv("SONIOX_API_KEY"),
settings=SonioxSTTService.Settings(
# Add language hints to use a specific language
# Add strict mode to enforce the language hints
language_hints=[Language.EN],
language_hints_strict=True,
),
)
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
task = PipelineTask(
@@ -109,7 +105,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

54
examples/foundational/07ab-interruptible-inworld-http.py → examples/foundational/07zb-interruptible-inworld-http.py
View File

@@ -10,9 +10,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame, TTSTextFrame
from pipecat.observers.loggers.debug_log_observer import DebugLogObserver, FrameEndpoint
from pipecat.pipeline.pipeline import Pipeline
@@ -23,7 +21,6 @@ from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.processors.frameworks.rtvi import RTVIObserver, RTVIProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.deepgram.stt import DeepgramSTTService
@@ -33,8 +30,6 @@ from pipecat.transports.base_output import BaseOutputTransport
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -42,17 +37,14 @@ transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -66,45 +58,36 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = InworldHttpTTSService(
api_key=os.getenv("INWORLD_API_KEY", ""),
aiohttp_session=session,
voice_id="Ashley",
model="inworld-tts-1",
# Set to False for non-streaming mode or True for streaming mode.
streaming=True,
settings=InworldHttpTTSService.Settings(
voice="Ashley",
model="inworld-tts-1",
),
# Set to False for non-streaming mode or True for streaming mode.
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful AI demonstrating Inworld AI's TTS. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a friendly and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful AI demonstrating Inworld AI's TTS. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a friendly and helpful way.",
),
)
rtvi = RTVIProcessor()
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(),
rtvi,
stt,
context_aggregator.user(),
user_aggregator,
llm,
tts,
transport.output(),
context_aggregator.assistant(),
assistant_aggregator,
]
)
@@ -115,7 +98,6 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
enable_usage_metrics=True,
),
observers=[
RTVIObserver(rtvi),
DebugLogObserver(
frame_types={
TTSTextFrame: (BaseOutputTransport, FrameEndpoint.SOURCE),
@@ -129,7 +111,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info("Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

65
examples/foundational/07ab-interruptible-inworld.py → examples/foundational/07zb-interruptible-inworld.py
View File

@@ -9,11 +9,8 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame, TTSTextFrame
from pipecat.observers.loggers.debug_log_observer import DebugLogObserver, FrameEndpoint
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
@@ -22,18 +19,14 @@ from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.processors.frameworks.rtvi import RTVIConfig, RTVIObserver, RTVIProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.inworld.tts import InworldTTSService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_output import BaseOutputTransport
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -42,17 +35,14 @@ transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -64,42 +54,35 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = InworldTTSService(
api_key=os.getenv("INWORLD_API_KEY", ""),
voice_id="Ashley",
model="inworld-tts-1",
temperature=1.1,
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful AI demonstrating Inworld AI's TTS. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a friendly and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
settings=InworldTTSService.Settings(
voice="Ashley",
model="inworld-tts-1",
temperature=1.1,
),
)
rtvi = RTVIProcessor(config=RTVIConfig(config=[]))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful AI demonstrating Inworld AI's TTS. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a friendly and helpful way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(),
rtvi,
stt,
context_aggregator.user(),
user_aggregator,
llm,
tts,
transport.output(),
context_aggregator.assistant(),
assistant_aggregator,
]
)
@@ -109,14 +92,6 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
enable_metrics=True,
enable_usage_metrics=True,
),
observers=[
RTVIObserver(rtvi),
DebugLogObserver(
frame_types={
TTSTextFrame: (BaseOutputTransport, FrameEndpoint.SOURCE),
}
),
],
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@@ -124,7 +99,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info("Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

52
examples/foundational/07ac-interruptible-asyncai-http.py → examples/foundational/07zc-interruptible-asyncai-http.py
View File

@@ -11,9 +11,7 @@ import aiohttp
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -31,30 +29,24 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -68,40 +60,34 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = AsyncAIHttpTTSService(
api_key=os.getenv("ASYNCAI_API_KEY", ""),
voice_id=os.getenv("ASYNCAI_VOICE_ID", "e0f39dc4-f691-4e78-bba5-5c636692cc04"),
settings=AsyncAIHttpTTSService.Settings(
voice="e0f39dc4-f691-4e78-bba5-5c636692cc04",
),
aiohttp_session=session,
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[
TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())
]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -118,7 +104,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message(
{"role": "user", "content": "Please introduce yourself to the user."}
)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

125
examples/foundational/07zc-interruptible-asyncai.py Normal file
View File

@@ -0,0 +1,125 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.asyncai.tts import AsyncAITTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = AsyncAITTSService(
api_key=os.getenv("ASYNCAI_API_KEY", ""),
settings=AsyncAITTSService.Settings(
voice="e0f39dc4-f691-4e78-bba5-5c636692cc04",
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

103
examples/foundational/07ad-interruptible-aicoustics.py → examples/foundational/07zd-interruptible-aicoustics.py
View File

@@ -13,7 +13,6 @@ from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.filters.aic_filter import AICFilter
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -32,56 +31,42 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# Create audio buffer processor so we can hear the audio fitler results.
audiobuffer = AudioBufferProcessor(
num_channels=2, # 1 for mono, 2 for stereo (user left, bot right)
enable_turn_audio=False, # Enable per-turn audio recording
)
def _create_aic_filter() -> AICFilter:
license_key = os.getenv("AICOUSTICS_LICENSE_KEY", "")
return AICFilter(
license_key=license_key,
enhancement_level=0.5,
model_id="quail-vf-2.0-l-16khz",
)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
aic_filter = _create_aic_filter()
aic_vad_analyzer = aic_filter.create_vad_analyzer(
speech_hold_duration=0.05, minimum_speech_duration=0.0, sensitivity=6.0
)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: (
lambda aic: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=aic.create_vad_analyzer(lookback_buffer_size=6.0, sensitivity=6.0),
audio_in_filter=aic,
)
)(_create_aic_filter()),
"twilio": lambda: (
lambda aic: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=aic.create_vad_analyzer(lookback_buffer_size=6.0, sensitivity=6.0),
audio_in_filter=aic,
)
)(_create_aic_filter()),
"webrtc": lambda: (
lambda aic: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=aic.create_vad_analyzer(lookback_buffer_size=6.0, sensitivity=6.0),
audio_in_filter=aic,
)
)(_create_aic_filter()),
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
audio_in_filter=aic_filter,
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
audio_in_filter=aic_filter,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
audio_in_filter=aic_filter,
),
}
@@ -92,38 +77,40 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=aic_vad_analyzer),
)
# Create audio buffer processor so we can hear the audio fitler results.
audiobuffer = AudioBufferProcessor(
num_channels=2, # 1 for mono, 2 for stereo (user left, bot right)
enable_turn_audio=False, # Enable per-turn audio recording
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
audiobuffer, # write audio data to a file
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -141,7 +128,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Client connected")
await audiobuffer.start_recording()
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@audiobuffer.event_handler("on_audio_data")

59
examples/foundational/07ae-interruptible-hume.py → examples/foundational/07ze-interruptible-hume.py
View File

@@ -9,9 +9,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame, TTSTextFrame
from pipecat.observers.loggers.debug_log_observer import DebugLogObserver, FrameEndpoint
from pipecat.pipeline.pipeline import Pipeline
@@ -22,7 +20,6 @@ from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.processors.frameworks.rtvi import RTVIConfig, RTVIObserver, RTVIProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.deepgram.stt import DeepgramSTTService
@@ -32,30 +29,24 @@ from pipecat.transports.base_output import BaseOutputTransport
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -68,40 +59,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = HumeTTSService(
api_key=os.getenv("HUME_API_KEY"),
# Replace with your Hume voice ID
voice_id="f898a92e-685f-43fa-985b-a46920f0650b",
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
settings=HumeTTSService.Settings(
voice="f898a92e-685f-43fa-985b-a46920f0650b",
),
)
rtvi = RTVIProcessor(config=RTVIConfig(config=[]))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
rtvi,
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS (HumeTTSService with word timestamps)
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -114,7 +98,6 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
observers=[
RTVIObserver(rtvi),
DebugLogObserver(
frame_types={
TTSTextFrame: (BaseOutputTransport, FrameEndpoint.SOURCE),
@@ -123,10 +106,6 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
],
)
@rtvi.event_handler("on_client_ready")
async def on_client_ready(rtvi):
await rtvi.set_bot_ready()
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
@@ -134,7 +113,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
"💡 Word timestamps are enabled! Watch the console for TTSTextFrame logs showing each word with its PTS."
)
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

58
examples/foundational/07af-interruptible-gradium.py → examples/foundational/07zf-interruptible-gradium.py
View File

@@ -9,9 +9,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -26,32 +24,27 @@ from pipecat.runner.utils import create_transport
from pipecat.services.gradium.stt import GradiumSTTService
from pipecat.services.gradium.tts import GradiumTTSService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -59,41 +52,44 @@ transport_params = {
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = GradiumSTTService(api_key=os.getenv("GRADIUM_API_KEY"))
stt = GradiumSTTService(
api_key=os.getenv("GRADIUM_API_KEY"),
api_endpoint_base_url="wss://us.api.gradium.ai/api/speech/asr",
settings=GradiumSTTService.Settings(
language=Language.EN,
),
)
tts = GradiumTTSService(
api_key=os.getenv("GRADIUM_API_KEY"),
voice_id="YTpq7expH9539ERJ",
url="wss://us.api.gradium.ai/api/speech/tts",
settings=GradiumTTSService.Settings(
voice="YTpq7expH9539ERJ",
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -110,7 +106,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

124
examples/foundational/07zg-interruptible-camb.py Normal file
View File

@@ -0,0 +1,124 @@
#
# Copyright (c) 2024–2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.camb.tts import CambTTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info("Starting Camb AI TTS bot")
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = CambTTSService(
api_key=os.getenv("CAMB_API_KEY"),
settings=CambTTSService.Settings(
model="mars-flash",
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful voice assistant powered by Camb AI text-to-speech. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Keep responses concise.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(),
stt,
user_aggregator,
llm,
tts,
transport.output(),
assistant_aggregator,
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
audio_out_sample_rate=22050,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info("Client connected")
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info("Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

50
examples/foundational/07ac-interruptible-asyncai.py → examples/foundational/07zi-interruptible-piper.py
View File

@@ -4,15 +4,12 @@
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -24,18 +21,15 @@ from pipecat.processors.aggregators.llm_response_universal import (
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.asyncai.tts import AsyncAITTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.piper.tts import PiperTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
@@ -43,17 +37,14 @@ transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -63,39 +54,34 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = AsyncAITTSService(
api_key=os.getenv("ASYNCAI_API_KEY", ""),
voice_id=os.getenv("ASYNCAI_VOICE_ID", "e0f39dc4-f691-4e78-bba5-5c636692cc04"),
tts = PiperTTSService(
settings=PiperTTSService.Settings(
voice="en_US-ryan-high",
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -112,7 +98,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

50
examples/foundational/07e-interruptible-playht-http.py → examples/foundational/07zj-interruptible-kokoro.py
View File

@@ -4,15 +4,12 @@
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -25,13 +22,11 @@ from pipecat.processors.aggregators.llm_response_universal import (
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.kokoro.tts import KokoroTTSService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.playht.tts import PlayHTHttpTTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -42,17 +37,14 @@ transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -62,40 +54,34 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = PlayHTHttpTTSService(
user_id=os.getenv("PLAYHT_USER_ID"),
api_key=os.getenv("PLAYHT_API_KEY"),
voice_url="s3://voice-cloning-zero-shot/d9ff78ba-d016-47f6-b0ef-dd630f59414e/female-cs/manifest.json",
tts = KokoroTTSService(
settings=KokoroTTSService.Settings(
voice="af_heart",
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -112,7 +98,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

123
examples/foundational/07zk-interruptible-resemble.py Normal file
View File

@@ -0,0 +1,123 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.services.resembleai.tts import ResembleAITTSService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = ResembleAITTSService(
api_key=os.getenv("RESEMBLE_API_KEY"),
settings=ResembleAITTSService.Settings(
voice=os.getenv("RESEMBLE_VOICE_UUID"),
),
)
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt,
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

47
examples/foundational/08-custom-frame-processor.py
View File

@@ -9,9 +9,7 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import (
Frame,
LLMRunFrame,
@@ -33,8 +31,6 @@ from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -77,20 +73,17 @@ class MetricsFrameLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
video_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -102,39 +95,35 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
metrics_frame_processor = MetricsFrameLogger()
pipeline = Pipeline(
[
transport.input(),
stt,
context_aggregator.user(),
user_aggregator,
llm,
tts,
transport.output(),
context_aggregator.assistant(),
assistant_aggregator,
metrics_frame_processor, # pretty print metrics frames
]
)
@@ -152,7 +141,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected: {client}")
# Kick off the conversation.
messages.append({"role": "system", "content": "Please introduce yourself to the user."})
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

5
examples/foundational/09-mirror.py
View File

@@ -47,9 +47,8 @@ class MirrorProcessor(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,

5
examples/foundational/09a-local-mirror.py
View File

@@ -50,9 +50,8 @@ class MirrorProcessor(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,

71
examples/foundational/10-wake-phrase.py
View File

@@ -9,10 +9,8 @@ import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import TTSSpeakFrame
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
@@ -21,7 +19,6 @@ from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.processors.filters.wake_check_filter import WakeCheckFilter
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.cartesia.tts import CartesiaTTSService
@@ -30,29 +27,28 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
from pipecat.turns.user_start import WakePhraseUserTurnStartStrategy
from pipecat.turns.user_turn_strategies import (
UserTurnStrategies,
default_user_turn_start_strategies,
)
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -60,44 +56,54 @@ transport_params = {
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
stt = DeepgramSTTService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
settings=DeepgramSTTService.Settings(
keyterm=["pipecat"],
),
)
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful assistant. Respond to what the user said in a creative and helpful way. Keep your responses brief.",
},
]
hey_robot_filter = WakeCheckFilter(["hey robot", "hey, robot"])
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
start=[
WakePhraseUserTurnStartStrategy(
phrases=["pipecat"],
# Timeout before wake phrase must be spoken again
timeout=5.0,
),
*default_user_turn_start_strategies(),
]
),
vad_analyzer=SileroVADAnalyzer(),
),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
hey_robot_filter, # Filter out speech not directed at the robot
context_aggregator.user(), # User responses
stt,
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -114,7 +120,8 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
async def on_client_connected(transport, client):
logger.info(f"Client connected")
# Kick off the conversation.
await task.queue_frame(TTSSpeakFrame("Hi! If you want to talk to me, just say 'Hey Robot'"))
context.add_message({"role": "user", "content": "Please introduce yourself to the user."})
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):

46
examples/foundational/11-sound-effects.py
View File

@@ -10,9 +10,7 @@ import wave
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import (
Frame,
LLMContextFrame,
@@ -22,7 +20,7 @@ from pipecat.frames.frames import (
)
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
@@ -38,8 +36,6 @@ from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
@@ -85,24 +81,20 @@ class InboundSoundEffectWrapper(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -112,28 +104,24 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way.",
),
)
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
out_sound = OutboundSoundEffectWrapper()
in_sound = InboundSoundEffectWrapper()
@@ -144,7 +132,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
[
transport.input(),
stt,
context_aggregator.user(),
user_aggregator,
in_sound,
fl2,
llm,
@@ -152,7 +140,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts,
out_sound,
transport.output(),
context_aggregator.assistant(),
assistant_aggregator,
]
)

139
examples/foundational/12-describe-image-openai-responses.py Normal file
View File

@@ -0,0 +1,139 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from PIL import Image
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.aggregators.llm_context import LLMContext
from pipecat.processors.aggregators.llm_response_universal import (
LLMContextAggregatorPair,
LLMUserAggregatorParams,
)
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.responses.llm import OpenAIResponsesLLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
load_dotenv(override=True)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = DeepgramSTTService(api_key=os.getenv("DEEPGRAM_API_KEY"))
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAIResponsesLLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAIResponsesLLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way. You are also able to describe images.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
assistant_aggregator, # Assistant spoken responses
]
)
task = PipelineTask(
pipeline,
params=PipelineParams(
enable_metrics=True,
enable_usage_metrics=True,
),
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_connected")
async def on_client_connected(transport, client):
logger.info(f"Client connected")
if not runner_args.body:
script_dir = os.path.dirname(__file__)
runner_args.body = {
"image_path": os.path.join(script_dir, "assets", "cat.jpg"),
"question": "Describe this image",
}
image_path = runner_args.body["image_path"]
question = runner_args.body["question"]
# Kick off the conversation.
image = Image.open(image_path)
message = await LLMContext.create_image_message(
image=image.tobytes(),
format="RGB",
size=image.size,
text=question,
)
context.add_message(message)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

47
examples/foundational/12-describe-image-openai.py
View File

@@ -11,9 +11,7 @@ from dotenv import load_dotenv
from loguru import logger
from PIL import Image
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,25 +28,20 @@ from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.openai.llm import OpenAILLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -60,37 +53,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = OpenAILLMService(api_key=os.getenv("OPENAI_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way. You are also able to describe images.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = OpenAILLMService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAILLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way. You are also able to describe images.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -125,7 +114,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
size=image.size,
text=question,
)
messages.append(message)
context.add_message(message)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

47
examples/foundational/12a-describe-image-anthropic.py
View File

@@ -11,9 +11,7 @@ from dotenv import load_dotenv
from loguru import logger
from PIL import Image
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,25 +28,20 @@ from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -60,37 +53,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = AnthropicLLMService(api_key=os.getenv("ANTHROPIC_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way. You are also able to describe images.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = AnthropicLLMService(
api_key=os.getenv("ANTHROPIC_API_KEY"),
settings=AnthropicLLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way. You are also able to describe images.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -125,7 +114,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
size=image.size,
text=question,
)
messages.append(message)
context.add_message(message)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

48
examples/foundational/12b-describe-image-aws.py
View File

@@ -11,9 +11,7 @@ from dotenv import load_dotenv
from loguru import logger
from PIL import Image
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,25 +28,20 @@ from pipecat.services.cartesia.tts import CartesiaTTSService
from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -60,44 +53,35 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = AWSBedrockLLMService(
aws_region="us-west-2",
model="us.anthropic.claude-3-7-sonnet-20250219-v1:0",
# Note: usually, prefer providing latency="optimized" param.
# Here we can't because AWS Bedrock doesn't support it for Claude 3.7,
# which we need for image input.
params=AWSBedrockLLMService.InputParams(temperature=0.8),
settings=AWSBedrockLLMService.Settings(
model="us.anthropic.claude-sonnet-4-6",
temperature=0.8,
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way. You are also able to describe images.",
),
)
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way. You are also able to describe images.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
),
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -132,7 +116,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
size=image.size,
text=question,
)
messages.append(message)
context.add_message(message)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

47
examples/foundational/12c-describe-image-gemini-flash.py
View File

@@ -11,9 +11,7 @@ from dotenv import load_dotenv
from loguru import logger
from PIL import Image
from pipecat.audio.turn.smart_turn.local_smart_turn_v3 import LocalSmartTurnAnalyzerV3
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import LLMRunFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -30,25 +28,20 @@ from pipecat.services.deepgram.stt import DeepgramSTTService
from pipecat.services.google.llm import GoogleLLMService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.turns.user_stop import TurnAnalyzerUserTurnStopStrategy
from pipecat.turns.user_turn_strategies import UserTurnStrategies
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -60,37 +53,33 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
llm = GoogleLLMService(api_key=os.getenv("GOOGLE_API_KEY"))
messages = [
{
"role": "system",
"content": "You are a helpful LLM in a WebRTC call. Your goal is to demonstrate your capabilities in a succinct way. Your output will be spoken aloud, so avoid special characters that can't easily be spoken, such as emojis or bullet points. Respond to what the user said in a creative and helpful way. You are also able to describe images.",
},
]
context = LLMContext(messages)
context_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(
user_turn_strategies=UserTurnStrategies(
stop=[TurnAnalyzerUserTurnStopStrategy(turn_analyzer=LocalSmartTurnAnalyzerV3())]
),
llm = GoogleLLMService(
api_key=os.getenv("GOOGLE_API_KEY"),
settings=GoogleLLMService.Settings(
system_instruction="You are a helpful assistant in a voice conversation. Your responses will be spoken aloud, so avoid emojis, bullet points, or other formatting that can't be spoken. Respond to what the user said in a creative, helpful, and brief way. You are also able to describe images.",
),
)
context = LLMContext()
user_aggregator, assistant_aggregator = LLMContextAggregatorPair(
context,
user_params=LLMUserAggregatorParams(vad_analyzer=SileroVADAnalyzer()),
)
pipeline = Pipeline(
[
transport.input(), # Transport user input
stt, # STT
context_aggregator.user(), # User responses
user_aggregator, # User responses
llm, # LLM
tts, # TTS
transport.output(), # Transport bot output
context_aggregator.assistant(), # Assistant spoken responses
assistant_aggregator, # Assistant spoken responses
]
)
@@ -125,7 +114,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
size=image.size,
text=question,
)
messages.append(message)
context.add_message(message)
await task.queue_frames([LLMRunFrame()])
@transport.event_handler("on_client_disconnected")

13
examples/foundational/12d-describe-image-moondream.py
View File

@@ -11,8 +11,6 @@ from dotenv import load_dotenv
from loguru import logger
from PIL import Image
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.audio.vad.vad_analyzer import VADParams
from pipecat.frames.frames import UserImageRawFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
@@ -27,17 +25,14 @@ from pipecat.transports.daily.transport import DailyParams
load_dotenv(override=True)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
"webrtc": lambda: TransportParams(
audio_out_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
),
}
@@ -47,7 +42,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
tts = CartesiaTTSService(
api_key=os.getenv("CARTESIA_API_KEY"),
voice_id="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
settings=CartesiaTTSService.Settings(
voice="71a7ad14-091c-4e8e-a314-022ece01c121", # British Reading Lady
),
)
vision = MoondreamService()

12
examples/foundational/13-whisper-transcription.py
View File

@@ -13,6 +13,7 @@ from pipecat.frames.frames import Frame, TranscriptionFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.audio.vad_processor import VADProcessor
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
@@ -35,21 +36,17 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
}
@@ -60,8 +57,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = WhisperSTTService()
tl = TranscriptionLogger()
vad_processor = VADProcessor(vad_analyzer=SileroVADAnalyzer())
pipeline = Pipeline([transport.input(), stt, tl])
pipeline = Pipeline([transport.input(), vad_processor, stt, tl])
task = PipelineTask(
pipeline,

5
examples/foundational/13a-whisper-local.py
View File

@@ -15,6 +15,7 @@ from pipecat.frames.frames import Frame, TranscriptionFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.audio.vad_processor import VADProcessor
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.services.whisper.stt import WhisperSTTService
from pipecat.transports.local.audio import LocalAudioTransport, LocalAudioTransportParams
@@ -40,15 +41,15 @@ async def main():
transport = LocalAudioTransport(
LocalAudioTransportParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
)
)
stt = WhisperSTTService()
tl = TranscriptionLogger()
vad_processor = VADProcessor(vad_analyzer=SileroVADAnalyzer())
pipeline = Pipeline([transport.input(), stt, tl])
pipeline = Pipeline([transport.input(), vad_processor, stt, tl])
task = PipelineTask(pipeline)

11
examples/foundational/13b-deepgram-transcription.py
View File

@@ -16,7 +16,7 @@ from pipecat.pipeline.task import PipelineTask
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.deepgram.stt import DeepgramSTTService, Language, LiveOptions
from pipecat.services.deepgram.stt import DeepgramSTTService, Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
@@ -35,9 +35,8 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),
@@ -50,7 +49,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = DeepgramSTTService(
api_key=os.getenv("DEEPGRAM_API_KEY"),
live_options=LiveOptions(language=Language.EN),
settings=DeepgramSTTService.Settings(
language=Language.EN,
),
)
tl = TranscriptionLogger()

11
examples/foundational/13c-gladia-transcription.py
View File

@@ -35,9 +35,8 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),
@@ -51,7 +50,11 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = GladiaSTTService(
api_key=os.getenv("GLADIA_API_KEY"),
region=os.getenv("GLADIA_REGION"),
# live_options=LiveOptions(language=Language.FR),
# settings=GladiaSTTSettings(
# language_config=LanguageConfig(
# languages=[Language.FR],
# ),
# ),
)
tl = TranscriptionLogger()

16
examples/foundational/13c-gladia-translation.py
View File

@@ -17,7 +17,6 @@ from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.gladia.config import (
GladiaInputParams,
LanguageConfig,
RealtimeProcessingConfig,
TranslationConfig,
@@ -44,9 +43,8 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),
@@ -60,16 +58,16 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = GladiaSTTService(
api_key=os.getenv("GLADIA_API_KEY"),
region=os.getenv("GLADIA_REGION"),
params=GladiaInputParams(
settings=GladiaSTTService.Settings(
language_config=LanguageConfig(
languages=[Language.EN], # Input in English
languages=[Language.EN],
code_switching=False,
),
realtime_processing=RealtimeProcessingConfig(
translation=True, # Enable translation
translation=True,
translation_config=TranslationConfig(
target_languages=[Language.ES], # Translate to Spanish
model="enhanced", # Use the enhanced translation model
target_languages=[Language.ES],
model="enhanced",
),
),
),

8
examples/foundational/13d-assemblyai-transcription.py
View File

@@ -35,9 +35,8 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),
@@ -50,6 +49,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = AssemblyAISTTService(
api_key=os.getenv("ASSEMBLYAI_API_KEY"),
settings=AssemblyAISTTService.Settings(
model="u3-rt-pro",
),
)
tl = TranscriptionLogger()

20
examples/foundational/13e-whisper-mlx.py
View File

@@ -15,6 +15,7 @@ from pipecat.frames.frames import Frame, TranscriptionFrame, UserStoppedSpeaking
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.audio.vad_processor import VADProcessor
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
@@ -56,21 +57,17 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=STOP_SECS)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=STOP_SECS)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=STOP_SECS)),
),
}
@@ -78,11 +75,18 @@ transport_params = {
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = WhisperSTTServiceMLX(model=MLXModel.LARGE_V3_TURBO)
stt = WhisperSTTServiceMLX(
settings=WhisperSTTServiceMLX.Settings(
model=MLXModel.LARGE_V3_TURBO.value,
),
)
tl = TranscriptionLogger()
vad_processor = VADProcessor(
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=STOP_SECS))
)
pipeline = Pipeline([transport.input(), stt, tl])
pipeline = Pipeline([transport.input(), vad_processor, stt, tl])
task = PipelineTask(
pipeline,

5
examples/foundational/13f-cartesia-transcription.py
View File

@@ -35,9 +35,8 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),

18
examples/foundational/13g-sambanova-transcription.py
View File

@@ -16,6 +16,7 @@ from pipecat.frames.frames import Frame, TranscriptionFrame, UserStoppedSpeaking
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineParams, PipelineTask
from pipecat.processors.audio.vad_processor import VADProcessor
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
@@ -57,21 +58,17 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=STOP_SECS)),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=STOP_SECS)),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=STOP_SECS)),
),
}
@@ -80,13 +77,18 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = SambaNovaSTTService(
model="Whisper-Large-v3",
settings=SambaNovaSTTService.Settings(
model="Whisper-Large-v3",
),
api_key=os.getenv("SAMBANOVA_API_KEY"),
)
tl = TranscriptionLogger()
vad_processor = VADProcessor(
vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=STOP_SECS))
)
pipeline = Pipeline([transport.input(), stt, tl])
pipeline = Pipeline([transport.input(), vad_processor, stt, tl])
task = PipelineTask(
pipeline,

7
examples/foundational/13h-speechmatics-transcription.py
View File

@@ -36,9 +36,8 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),
@@ -66,7 +65,7 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
stt = SpeechmaticsSTTService(
api_key=os.getenv("SPEECHMATICS_API_KEY"),
params=SpeechmaticsSTTService.InputParams(
settings=SpeechmaticsSTTService.Settings(
language=Language.EN,
speaker_active_format="<{speaker_id}>{text}</{speaker_id}>",
),

7
examples/foundational/13i-soniox-transcription.py
View File

@@ -14,6 +14,7 @@ from pipecat.frames.frames import Frame, TranscriptionFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.audio.vad_processor import VADProcessor
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
@@ -39,15 +40,12 @@ class TranscriptionLogger(FrameProcessor):
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
}
@@ -60,8 +58,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
)
tl = TranscriptionLogger()
vad_processor = VADProcessor(vad_analyzer=SileroVADAnalyzer())
pipeline = Pipeline([transport.input(), stt, tl])
pipeline = Pipeline([transport.input(), vad_processor, stt, tl])
task = PipelineTask(
pipeline,

7
examples/foundational/13j-azure-transcription.py
View File

@@ -14,6 +14,7 @@ from pipecat.frames.frames import Frame, TranscriptionFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.audio.vad_processor import VADProcessor
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
@@ -39,15 +40,12 @@ class TranscriptionLogger(FrameProcessor):
transport_params = {
"daily": lambda: DailyParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
"webrtc": lambda: TransportParams(
audio_in_enabled=True,
vad_analyzer=SileroVADAnalyzer(),
),
}
@@ -61,8 +59,9 @@ async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
)
tl = TranscriptionLogger()
vad_processor = VADProcessor(vad_analyzer=SileroVADAnalyzer())
pipeline = Pipeline([transport.input(), stt, tl])
pipeline = Pipeline([transport.input(), vad_processor, stt, tl])
task = PipelineTask(
pipeline,

48
examples/foundational/13k-elevenlabs-transcription.py
View File

@@ -6,7 +6,6 @@
import os
import aiohttp
from dotenv import load_dotenv
from loguru import logger
@@ -15,10 +14,11 @@ from pipecat.frames.frames import Frame, TranscriptionFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.audio.vad_processor import VADProcessor
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.elevenlabs.stt import ElevenLabsSTTService
from pipecat.services.elevenlabs.stt import ElevenLabsRealtimeSTTService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
@@ -37,44 +37,38 @@ class TranscriptionLogger(FrameProcessor):
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True, vad_analyzer=SileroVADAnalyzer()),
"twilio": lambda: FastAPIWebsocketParams(
audio_in_enabled=True, vad_analyzer=SileroVADAnalyzer()
),
"webrtc": lambda: TransportParams(audio_in_enabled=True, vad_analyzer=SileroVADAnalyzer()),
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),
"webrtc": lambda: TransportParams(audio_in_enabled=True),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
async with aiohttp.ClientSession() as session:
stt = ElevenLabsSTTService(
api_key=os.getenv("ELEVENLABS_API_KEY"),
aiohttp_session=session,
)
stt = ElevenLabsRealtimeSTTService(api_key=os.getenv("ELEVENLABS_API_KEY"))
tl = TranscriptionLogger()
tl = TranscriptionLogger()
vad_processor = VADProcessor(vad_analyzer=SileroVADAnalyzer())
pipeline = Pipeline([transport.input(), stt, tl])
pipeline = Pipeline([transport.input(), vad_processor, stt, tl])
task = PipelineTask(
pipeline,
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
task = PipelineTask(
pipeline,
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
await runner.run(task)
async def bot(runner_args: RunnerArguments):

89
examples/foundational/13l-gradium-transcription.py Normal file
View File

@@ -0,0 +1,89 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.frames.frames import Frame, TranscriptionFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.gradium.stt import GradiumSTTService
from pipecat.transcriptions.language import Language
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
class TranscriptionLogger(FrameProcessor):
async def process_frame(self, frame: Frame, direction: FrameDirection):
await super().process_frame(frame, direction)
if isinstance(frame, TranscriptionFrame):
print(f"Transcription: {frame.text}")
# Push all frames through
await self.push_frame(frame, direction)
# We store functions so objects (e.g. SileroVADAnalyzer) don't get
# instantiated. The function will be called when the desired transport gets
# selected.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),
"webrtc": lambda: TransportParams(audio_in_enabled=True),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = GradiumSTTService(
api_key=os.getenv("GRADIUM_API_KEY"),
api_endpoint_base_url="wss://us.api.gradium.ai/api/speech/asr",
settings=GradiumSTTService.Settings(
language=Language.EN,
delay_in_frames=8,
),
)
tl = TranscriptionLogger()
pipeline = Pipeline([transport.input(), stt, tl])
task = PipelineTask(
pipeline,
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

89
examples/foundational/13m-openai-transcription.py Normal file
View File

@@ -0,0 +1,89 @@
#
# Copyright (c) 2024-2026, Daily
#
# SPDX-License-Identifier: BSD 2-Clause License
#
import os
from dotenv import load_dotenv
from loguru import logger
from pipecat.audio.vad.silero import SileroVADAnalyzer
from pipecat.frames.frames import Frame, TranscriptionFrame
from pipecat.pipeline.pipeline import Pipeline
from pipecat.pipeline.runner import PipelineRunner
from pipecat.pipeline.task import PipelineTask
from pipecat.processors.audio.vad_processor import VADProcessor
from pipecat.processors.frame_processor import FrameDirection, FrameProcessor
from pipecat.runner.types import RunnerArguments
from pipecat.runner.utils import create_transport
from pipecat.services.openai.stt import OpenAIRealtimeSTTService
from pipecat.transports.base_transport import BaseTransport, TransportParams
from pipecat.transports.daily.transport import DailyParams
from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
load_dotenv(override=True)
class TranscriptionLogger(FrameProcessor):
async def process_frame(self, frame: Frame, direction: FrameDirection):
await super().process_frame(frame, direction)
if isinstance(frame, TranscriptionFrame):
print(f"Transcription: {frame.text}")
# Push all frames through
await self.push_frame(frame, direction)
# We use lambdas to defer transport parameter creation until the transport
# type is selected at runtime.
transport_params = {
"daily": lambda: DailyParams(audio_in_enabled=True),
"twilio": lambda: FastAPIWebsocketParams(audio_in_enabled=True),
"webrtc": lambda: TransportParams(audio_in_enabled=True),
}
async def run_bot(transport: BaseTransport, runner_args: RunnerArguments):
logger.info(f"Starting bot")
stt = OpenAIRealtimeSTTService(
api_key=os.getenv("OPENAI_API_KEY"),
settings=OpenAIRealtimeSTTService.Settings(
model="gpt-4o-transcribe",
prompt="Expect words related to dogs, such as breed names.",
),
)
tl = TranscriptionLogger()
vad_processor = VADProcessor(vad_analyzer=SileroVADAnalyzer())
pipeline = Pipeline([transport.input(), vad_processor, stt, tl])
task = PipelineTask(
pipeline,
idle_timeout_secs=runner_args.pipeline_idle_timeout_secs,
)
@transport.event_handler("on_client_disconnected")
async def on_client_disconnected(transport, client):
logger.info(f"Client disconnected")
await task.cancel()
runner = PipelineRunner(handle_sigint=runner_args.handle_sigint)
await runner.run(task)
async def bot(runner_args: RunnerArguments):
"""Main bot entry point compatible with Pipecat Cloud."""
transport = await create_transport(runner_args, transport_params)
await run_bot(transport, runner_args)
if __name__ == "__main__":
from pipecat.runner.run import main
main()

Some files were not shown because too many files have changed in this diff Show More