diff --git a/CHANGELOG.md b/CHANGELOG.md index ae7e3dcb1..0ff7a6118 100644 --- a/CHANGELOG.md +++ b/CHANGELOG.md @@ -9,6 +9,9 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0 ### Added +- Added support for Nano Banana models to `GoogleLLMService`. For example, you + can now use the `gemini-2.5-flash-image` model to generate images. + - `PermissionError` is now caught if NLTK's `punkt_tab` can't be downloaded. - Added `HumeTTSService` for text-to-speech synthesis using Hume AI's diff --git a/examples/foundational/07n-interruptible-gemini-image.py b/examples/foundational/07n-interruptible-gemini-image.py new file mode 100644 index 000000000..61b8e650a --- /dev/null +++ b/examples/foundational/07n-interruptible-gemini-image.py @@ -0,0 +1,151 @@ +# +# Copyright (c) 2024–2025, Daily +# +# SPDX-License-Identifier: BSD 2-Clause License +# + +""" +A conversational AI bot using Gemini for both LLM, STT and TTS. + +This example demonstrates how to use Gemini's image generation capabilities. + +Features showcased: +- Gemini LLM for conversation and image generation +- Google TTS and STT + +Run with: + python examples/foundational/07n-interruptible-gemini-image.py + +Make sure to set your environment variables: + export GOOGLE_API_KEY=your_api_key_here +""" + +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 +from pipecat.pipeline.task import PipelineParams, PipelineTask +from pipecat.processors.aggregators.llm_context import LLMContext +from pipecat.processors.aggregators.llm_response_universal import LLMContextAggregatorPair +from pipecat.runner.types import RunnerArguments +from pipecat.runner.utils import create_transport +from pipecat.services.google.llm import GoogleLLMService +from pipecat.services.google.stt import GoogleSTTService +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 + +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. +transport_params = { + "daily": lambda: DailyParams( + audio_in_enabled=True, + audio_out_enabled=True, + video_out_enabled=True, + video_out_width=1024, + video_out_height=1024, + vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)), + turn_analyzer=LocalSmartTurnAnalyzerV3(params=SmartTurnParams()), + ), + "webrtc": lambda: TransportParams( + audio_in_enabled=True, + audio_out_enabled=True, + video_out_enabled=True, + video_out_width=1024, + video_out_height=1024, + vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)), + turn_analyzer=LocalSmartTurnAnalyzerV3(params=SmartTurnParams()), + ), +} + + +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"), + ) + + tts = GoogleTTSService( + voice_id="en-US-Chirp3-HD-Charon", + params=GoogleTTSService.InputParams(language=Language.EN_US), + credentials=os.getenv("GOOGLE_TEST_CREDENTIALS"), + ) + + llm = GoogleLLMService( + api_key=os.getenv("GOOGLE_API_KEY"), + model="gemini-2.5-flash-image", + ) + + 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 converted to audio so don't include special characters in your answers. Respond to what the user said in a creative and helpful way.", + }, + ] + + context = LLMContext(messages) + context_aggregator = LLMContextAggregatorPair(context) + + pipeline = Pipeline( + [ + transport.input(), # Transport user input + stt, # STT + context_aggregator.user(), # User responses + llm, # LLM + tts, # Gemini TTS + transport.output(), # Transport bot output + context_aggregator.assistant(), # 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 with a styled introduction + messages.append({"role": "system", "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() diff --git a/src/pipecat/services/google/llm.py b/src/pipecat/services/google/llm.py index 70d4ca2bf..b7b3bb50c 100644 --- a/src/pipecat/services/google/llm.py +++ b/src/pipecat/services/google/llm.py @@ -35,6 +35,7 @@ from pipecat.frames.frames import ( LLMMessagesFrame, LLMTextFrame, LLMUpdateSettingsFrame, + OutputImageRawFrame, UserImageRawFrame, ) from pipecat.metrics.metrics import LLMTokenUsage @@ -72,6 +73,9 @@ try: HttpOptions, Part, ) + + # Temporary hack to be able to process Nano Banana returned images. + genai._api_client.READ_BUFFER_SIZE = 5 * 1024 * 1024 except ModuleNotFoundError as e: logger.error(f"Exception: {e}") logger.error("In order to use Google AI, you need to `pip install pipecat-ai[google]`.") @@ -710,6 +714,7 @@ class GoogleLLMService(LLMService): self._api_key = api_key self._system_instruction = system_instruction self._http_options = http_options + self._create_client(api_key, http_options) self._settings = { "max_tokens": params.max_tokens, @@ -788,6 +793,9 @@ class GoogleLLMService(LLMService): # and can be configured to turn it off. if not self._model_name.startswith("gemini-2.5-flash"): return + # If we have an image model, we don't use a budget either. + if "image" in self._model_name: + return # If thinking_config is already set, don't override it. if "thinking_config" in generation_params: return @@ -927,6 +935,12 @@ class GoogleLLMService(LLMService): arguments=function_call.args or {}, ) ) + elif part.inline_data and part.inline_data.data: + image = Image.open(io.BytesIO(part.inline_data.data)) + frame = OutputImageRawFrame( + image=image.tobytes(), size=image.size, format="RGB" + ) + await self.push_frame(frame) if ( candidate.grounding_metadata