Final fix for AsyncAI TTS transcription duplication

The debug logging confirmed the issue was resolved with the proper flag tracking
and reset_word_timestamps() call. AsyncAI TTS now generates transcription frames
once per text input, matching Cartesia TTS behavior.

Working solution:
- Track transcription generation with _transcription_generated flag
- Reset flag and word timestamps for each new text input
- Only generate transcription on first audio chunk per text
- Proper cleanup ensures no duplicate transcriptions

CHANGELOG.md

@@ -19,6 +19,23 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
       ...
   ```
 
+### Changed
+
+- Update `livekit` to 1.0.13.
+
+- `torch` and `torchaudio` are no longer required for running Smart Turn
+  locally. This avoids gigabytes of dependencies being installed.
+
+### Changed
+
+- Updated `websockets` dependency to support version 15.0. Removed deprecated
+  usage of `ConnectionClosed.code` and `ConnectionClosed.reason` attributes in
+  `AWSTranscribeSTTService` for compatibility.
+
+- Refactored `pyproject.toml` to reduce websockets dependency repetition using
+  self-referencing extras. All websockets-dependent services now reference a
+  shared `websockets-base` extra.
+
 ### Deprecated
 
 - `PipelineTask` events `on_pipeline_stopped`, `on_pipeline_ended` and
@@ -27,6 +44,10 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ### Fixed
 
+- Fixed a self-cancellation deadlock in `UserIdleProcessor` when returning
+  `False` from an idle callback. The task now terminates naturally instead of
+  attempting to cancel itself.
+
 - Fixed an issue in `AudioBufferProcessor` where a recording is not created
   when a bot speaks and user input is blocked.
 

examples/foundational/01b-livekit-audio.py

@@ -11,7 +11,7 @@ import sys
 from dotenv import load_dotenv
 from loguru import logger
 
-from pipecat.frames.frames import TextFrame
+from pipecat.frames.frames import TTSSpeakFrame
 from pipecat.pipeline.pipeline import Pipeline
 from pipecat.pipeline.runner import PipelineRunner
 from pipecat.pipeline.task import PipelineTask
@@ -50,7 +50,7 @@ async def main():
     async def on_first_participant_joined(transport, participant_id):
         await asyncio.sleep(1)
         await task.queue_frame(
-            TextFrame(
+            TTSSpeakFrame(
                 "Hello there! How are you doing today? Would you like to talk about the weather?"
             )
         )

examples/foundational/38b-smart-turn-local.py

@@ -30,10 +30,6 @@ from pipecat.transports.websocket.fastapi import FastAPIWebsocketParams
 
 load_dotenv(override=True)
 
-# To use this locally, set the environment variable LOCAL_SMART_TURN_MODEL_PATH
-# to the Smart Turn v3 ONNX model file.
-smart_turn_model_path = os.getenv("LOCAL_SMART_TURN_MODEL_PATH")
-
 # We store functions so objects (e.g. SileroVADAnalyzer) don't get
 # instantiated. The function will be called when the desired transport gets
 # selected.
@@ -42,25 +38,19 @@ transport_params = {
         audio_in_enabled=True,
         audio_out_enabled=True,
         vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
-        turn_analyzer=LocalSmartTurnAnalyzerV3(
-            smart_turn_model_path=smart_turn_model_path, params=SmartTurnParams()
-        ),
+        turn_analyzer=LocalSmartTurnAnalyzerV3(params=SmartTurnParams()),
     ),
     "twilio": lambda: FastAPIWebsocketParams(
         audio_in_enabled=True,
         audio_out_enabled=True,
         vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
-        turn_analyzer=LocalSmartTurnAnalyzerV3(
-            smart_turn_model_path=smart_turn_model_path, params=SmartTurnParams()
-        ),
+        turn_analyzer=LocalSmartTurnAnalyzerV3(params=SmartTurnParams()),
     ),
     "webrtc": lambda: TransportParams(
         audio_in_enabled=True,
         audio_out_enabled=True,
         vad_analyzer=SileroVADAnalyzer(params=VADParams(stop_secs=0.2)),
-        turn_analyzer=LocalSmartTurnAnalyzerV3(
-            smart_turn_model_path=smart_turn_model_path, params=SmartTurnParams()
-        ),
+        turn_analyzer=LocalSmartTurnAnalyzerV3(params=SmartTurnParams()),
     ),
 }
 

pyproject.toml

@@ -47,32 +47,32 @@ Website = "https://pipecat.ai"
 [project.optional-dependencies]
 aic = [ "aic-sdk~=1.0.1" ]
 anthropic = [ "anthropic~=0.49.0" ]
-assemblyai = [ "websockets>=13.1,<15.0" ]
-asyncai = [ "websockets>=13.1,<15.0" ]
-aws = [ "aioboto3~=15.0.0", "websockets>=13.1,<15.0" ]
+assemblyai = [ "pipecat-ai[websockets-base]" ]
+asyncai = [ "pipecat-ai[websockets-base]" ]
+aws = [ "aioboto3~=15.0.0", "pipecat-ai[websockets-base]" ]
 aws-nova-sonic = [ "aws_sdk_bedrock_runtime~=0.0.2; python_version>='3.12'" ]
 azure = [ "azure-cognitiveservices-speech~=1.42.0"]
-cartesia = [ "cartesia~=2.0.3", "websockets>=13.1,<15.0" ]
+cartesia = [ "cartesia~=2.0.3", "pipecat-ai[websockets-base]" ]
 cerebras = []
 deepseek = []
 daily = [ "daily-python~=0.19.9" ]
 deepgram = [ "deepgram-sdk~=4.7.0" ]
-elevenlabs = [ "websockets>=13.1,<15.0" ]
+elevenlabs = [ "pipecat-ai[websockets-base]" ]
 fal = [ "fal-client~=0.5.9" ]
 fireworks = []
-fish = [ "ormsgpack~=1.7.0", "websockets>=13.1,<15.0" ]
-gladia = [ "websockets>=13.1,<15.0" ]
-google = [ "google-cloud-speech~=2.32.0", "google-cloud-texttospeech~=2.26.0", "google-genai~=1.24.0", "websockets>=13.1,<15.0" ]
+fish = [ "ormsgpack~=1.7.0", "pipecat-ai[websockets-base]" ]
+gladia = [ "pipecat-ai[websockets-base]" ]
+google = [ "google-cloud-speech~=2.32.0", "google-cloud-texttospeech~=2.26.0", "google-genai~=1.24.0", "pipecat-ai[websockets-base]" ]
 grok = []
 groq = [ "groq~=0.23.0" ]
 gstreamer = [ "pygobject~=3.50.0" ]
-heygen = [ "livekit>=0.22.0", "websockets>=13.1,<15.0" ]
+heygen = [ "livekit>=1.0.13", "pipecat-ai[websockets-base]" ]
 inworld = []
 krisp = [ "pipecat-ai-krisp~=0.4.0" ]
 koala = [ "pvkoala~=2.0.3" ]
 langchain = [ "langchain~=0.3.20", "langchain-community~=0.3.20", "langchain-openai~=0.3.9" ]
-livekit = [ "livekit~=0.22.0", "livekit-api~=0.8.2", "tenacity>=8.2.3,<10.0.0" ]
-lmnt = [ "websockets>=13.1,<15.0" ]
+livekit = [ "livekit~=1.0.13", "livekit-api~=1.0.5", "tenacity>=8.2.3,<10.0.0" ]
+lmnt = [ "pipecat-ai[websockets-base]" ]
 local = [ "pyaudio~=0.2.14" ]
 mcp = [ "mcp[cli]~=1.9.4" ]
 mem0 = [ "mem0ai~=0.1.94" ]
@@ -80,26 +80,26 @@ mistral = []
 mlx-whisper = [ "mlx-whisper~=0.4.2" ]
 moondream = [ "accelerate~=1.10.0", "einops~=0.8.0", "pyvips[binary]~=3.0.0", "timm~=1.0.13", "transformers>=4.48.0" ]
 nim = []
-neuphonic = [ "websockets>=13.1,<15.0" ]
+neuphonic = [ "pipecat-ai[websockets-base]" ]
 noisereduce = [ "noisereduce~=3.0.3" ]
-openai = [ "websockets>=13.1,<15.0" ]
+openai = [ "pipecat-ai[websockets-base]" ]
 openpipe = [ "openpipe~=4.50.0" ]
 openrouter = []
 perplexity = []
-playht = [ "websockets>=13.1,<15.0" ]
+playht = [ "pipecat-ai[websockets-base]" ]
 qwen = []
-rime = [ "websockets>=13.1,<15.0" ]
+rime = [ "pipecat-ai[websockets-base]" ]
 riva = [ "nvidia-riva-client~=2.21.1" ]
 runner = [ "python-dotenv>=1.0.0,<2.0.0", "uvicorn>=0.32.0,<1.0.0", "fastapi>=0.115.6,<0.117.0", "pipecat-ai-small-webrtc-prebuilt>=1.0.0"]
 sambanova = []
-sarvam = [ "websockets>=13.1,<15.0" ]
+sarvam = [ "pipecat-ai[websockets-base]" ]
 sentry = [ "sentry-sdk~=2.23.1" ]
 local-smart-turn = [ "coremltools>=8.0", "transformers", "torch>=2.5.0,<3", "torchaudio>=2.5.0,<3" ]
-local-smart-turn-v3 = [ "transformers", "torch>=2.5.0,<3", "torchaudio>=2.5.0,<3", "onnxruntime>=1.20.1, <2" ]
+local-smart-turn-v3 = [ "transformers", "onnxruntime>=1.20.1, <2" ]
 remote-smart-turn = []
 silero = [ "onnxruntime>=1.20.1, <2" ]
 simli = [ "simli-ai~=0.1.10"]
-soniox = [ "websockets>=13.1,<15.0" ]
+soniox = [ "pipecat-ai[websockets-base]" ]
 soundfile = [ "soundfile~=0.13.0" ]
 speechmatics = [ "speechmatics-rt>=0.4.0" ]
 tavus=[]
@@ -107,7 +107,8 @@ together = []
 tracing = [ "opentelemetry-sdk>=1.33.0", "opentelemetry-api>=1.33.0", "opentelemetry-instrumentation>=0.54b0" ]
 ultravox = [ "transformers>=4.48.0", "vllm>=0.9.0" ]
 webrtc = [ "aiortc~=1.11.0", "opencv-python~=4.11.0.86" ]
-websocket = [ "websockets>=13.1,<15.0", "fastapi>=0.115.6,<0.117.0" ]
+websocket = [ "pipecat-ai[websockets-base]", "fastapi>=0.115.6,<0.117.0" ]
+websockets-base = [ "websockets>=13.1,<16.0" ]
 whisper = [ "faster-whisper~=1.1.1" ]
 
 [dependency-groups]

src/pipecat/audio/turn/smart_turn/local_smart_turn_v3.py

@@ -98,15 +98,15 @@ class LocalSmartTurnAnalyzerV3(BaseSmartTurn):
         inputs = self._feature_extractor(
             audio_array,
             sampling_rate=16000,
-            return_tensors="pt",
+            return_tensors="np",
             padding="max_length",
             max_length=8 * 16000,
             truncation=True,
             do_normalize=True,
         )
 
-        # Convert to numpy and ensure correct shape for ONNX
-        input_features = inputs.input_features.squeeze(0).numpy().astype(np.float32)
+        # Extract features and ensure correct shape for ONNX
+        input_features = inputs.input_features.squeeze(0).astype(np.float32)
         input_features = np.expand_dims(input_features, axis=0)  # Add batch dimension
 
         # Run ONNX inference

src/pipecat/processors/user_idle_processor.py

@@ -17,7 +17,6 @@ from pipecat.frames.frames import (
     Frame,
     FunctionCallInProgressFrame,
     FunctionCallResultFrame,
-    StartFrame,
     UserStartedSpeakingFrame,
     UserStoppedSpeakingFrame,
 )
@@ -185,15 +184,13 @@ class UserIdleProcessor(FrameProcessor):
 
         Runs in a loop until cancelled or callback indicates completion.
         """
-        while True:
+        running = True
+        while running:
             try:
                 await asyncio.wait_for(self._idle_event.wait(), timeout=self._timeout)
             except asyncio.TimeoutError:
                 if not self._interrupted:
                     self._retry_count += 1
-                    should_continue = await self._callback(self, self._retry_count)
-                    if not should_continue:
-                        await self._stop()
-                        break
+                    running = await self._callback(self, self._retry_count)
             finally:
                 self._idle_event.clear()

src/pipecat/runner/run.py

@@ -70,7 +70,6 @@ import asyncio
 import os
 import sys
 from contextlib import asynccontextmanager
-from typing import Dict
 
 from loguru import logger
 
@@ -79,6 +78,10 @@ from pipecat.runner.types import (
     SmallWebRTCRunnerArguments,
     WebSocketRunnerArguments,
 )
+from pipecat.transports.smallwebrtc.request_handler import (
+    SmallWebRTCRequest,
+    SmallWebRTCRequestHandler,
+)
 
 try:
     import uvicorn
@@ -187,9 +190,6 @@ def _setup_webrtc_routes(app: FastAPI, esp32_mode: bool = False, host: str = "lo
         logger.error(f"WebRTC transport dependencies not installed: {e}")
         return
 
-    # Store connections by pc_id
-    pcs_map: Dict[str, SmallWebRTCConnection] = {}
-
     # Mount the frontend
     app.mount("/client", SmallWebRTCPrebuiltUI)
 
@@ -198,51 +198,33 @@ def _setup_webrtc_routes(app: FastAPI, esp32_mode: bool = False, host: str = "lo
         """Redirect root requests to client interface."""
         return RedirectResponse(url="/client/")
 
+    # Initialize the SmallWebRTC request handler
+    small_webrtc_handler: SmallWebRTCRequestHandler = SmallWebRTCRequestHandler(
+        esp32_mode=esp32_mode, host=host
+    )
+
     @app.post("/api/offer")
-    async def offer(request: dict, background_tasks: BackgroundTasks):
-        """Handle WebRTC offer requests and manage peer connections."""
-        pc_id = request.get("pc_id")
-
-        if pc_id and pc_id in pcs_map:
-            pipecat_connection = pcs_map[pc_id]
-            logger.info(f"Reusing existing connection for pc_id: {pc_id}")
-            await pipecat_connection.renegotiate(
-                sdp=request["sdp"],
-                type=request["type"],
-                restart_pc=request.get("restart_pc", False),
-            )
-        else:
-            pipecat_connection = SmallWebRTCConnection()
-            await pipecat_connection.initialize(sdp=request["sdp"], type=request["type"])
-
-            @pipecat_connection.event_handler("closed")
-            async def handle_disconnected(webrtc_connection: SmallWebRTCConnection):
-                """Handle WebRTC connection closure and cleanup."""
-                logger.info(f"Discarding peer connection for pc_id: {webrtc_connection.pc_id}")
-                pcs_map.pop(webrtc_connection.pc_id, None)
+    async def offer(request: SmallWebRTCRequest, background_tasks: BackgroundTasks):
+        """Handle WebRTC offer requests via SmallWebRTCRequestHandler."""
 
+        # Prepare runner arguments with the callback to run your bot
+        async def webrtc_connection_callback(connection):
             bot_module = _get_bot_module()
-            runner_args = SmallWebRTCRunnerArguments(webrtc_connection=pipecat_connection)
+            runner_args = SmallWebRTCRunnerArguments(webrtc_connection=connection)
             background_tasks.add_task(bot_module.bot, runner_args)
 
-        answer = pipecat_connection.get_answer()
-
-        # Apply ESP32 SDP munging if enabled
-        if esp32_mode and host != "localhost":
-            from pipecat.runner.utils import smallwebrtc_sdp_munging
-
-            answer["sdp"] = smallwebrtc_sdp_munging(answer["sdp"], host)
-
-        pcs_map[answer["pc_id"]] = pipecat_connection
+        # Delegate handling to SmallWebRTCRequestHandler
+        answer = await small_webrtc_handler.handle_web_request(
+            request=request,
+            webrtc_connection_callback=webrtc_connection_callback,
+        )
         return answer
 
     @asynccontextmanager
     async def lifespan(app: FastAPI):
         """Manage FastAPI application lifecycle and cleanup connections."""
         yield
-        coros = [pc.disconnect() for pc in pcs_map.values()]
-        await asyncio.gather(*coros)
-        pcs_map.clear()
+        await small_webrtc_handler.close()
 
     app.router.lifespan_context = lifespan
 

src/pipecat/services/asyncai/tts.py

@@ -27,7 +27,7 @@ from pipecat.frames.frames import (
     TTSStoppedFrame,
 )
 from pipecat.processors.frame_processor import FrameDirection
-from pipecat.services.tts_service import InterruptibleTTSService, TTSService
+from pipecat.services.tts_service import TTSService, WebsocketWordTTSService
 from pipecat.transcriptions.language import Language
 from pipecat.utils.tracing.service_decorators import traced_tts
 
@@ -71,7 +71,7 @@ def language_to_async_language(language: Language) -> Optional[str]:
     return result
 
 
-class AsyncAITTSService(InterruptibleTTSService):
+class AsyncAITTSService(WebsocketWordTTSService):
     """Async TTS service with WebSocket streaming.
 
     Provides text-to-speech using Async's streaming WebSocket API.
@@ -148,6 +148,8 @@ class AsyncAITTSService(InterruptibleTTSService):
         self._receive_task = None
         self._keepalive_task = None
         self._started = False
+        self._current_text = ""  # Track current text for generating timestamps
+        self._transcription_generated = False  # Track if transcription was already generated
 
     def can_generate_metrics(self) -> bool:
         """Check if this service can generate processing metrics.
@@ -286,12 +288,28 @@ class AsyncAITTSService(InterruptibleTTSService):
 
             elif msg.get("audio"):
                 await self.stop_ttfb_metrics()
+
+                # Start word timestamps and add the entire text as one "word"
+                # This generates transcription frames for the bot - but only once per text
+                if not self._transcription_generated and self._current_text.strip():
+                    logger.debug(
+                        f"{self}: Generating transcription for text: '{self._current_text}'"
+                    )
+                    self.start_word_timestamps()
+                    # Add the entire text as a single timestamp at time 0
+                    await self.add_word_timestamps([(self._current_text, 0.0)])
+                    # Add stop markers to end the word timestamps
+                    await self.add_word_timestamps([("TTSStoppedFrame", 0), ("Reset", 0)])
+                    self._transcription_generated = True
+                    logger.debug(f"{self}: Transcription flag set to True")
+
                 frame = TTSAudioRawFrame(
                     audio=base64.b64decode(msg["audio"]),
                     sample_rate=self.sample_rate,
                     num_channels=1,
                 )
                 await self.push_frame(frame)
+
             elif msg.get("error_code"):
                 logger.error(f"{self} error: {msg}")
                 await self.push_frame(TTSStoppedFrame())
@@ -324,7 +342,7 @@ class AsyncAITTSService(InterruptibleTTSService):
         Yields:
             Frame: Audio frames containing the synthesized speech.
         """
-        logger.debug(f"{self}: Generating TTS [{text}]")
+        logger.debug(f"{self}: AsyncAI Generating TTS [{text}]")
 
         try:
             if not self._websocket or self._websocket.state is State.CLOSED:
@@ -335,6 +353,12 @@ class AsyncAITTSService(InterruptibleTTSService):
                 yield TTSStartedFrame()
                 self._started = True
 
+            # Store the current text for generating timestamps
+            self._current_text = text
+            self._transcription_generated = False  # Reset for new text
+            # Reset word timestamps for new text to ensure clean state
+            self.reset_word_timestamps()
+            logger.debug(f"{self}: New text set: '{text}', transcription flag reset to False")
             msg = self._build_msg(text=text, force=True)
 
             try:
@@ -346,7 +370,6 @@ class AsyncAITTSService(InterruptibleTTSService):
                 await self._disconnect()
                 await self._connect()
                 return
-            yield None
         except Exception as e:
             logger.error(f"{self} exception: {e}")
 

src/pipecat/services/aws/stt.py

@@ -532,9 +532,7 @@ class AWSTranscribeSTTService(STTService):
                     logger.debug(f"{self} Other message type received: {headers}")
                     logger.debug(f"{self} Payload: {payload}")
             except websockets.exceptions.ConnectionClosed as e:
-                logger.error(
-                    f"{self} WebSocket connection closed in receive loop with code {e.code}: {e.reason}"
-                )
+                logger.error(f"{self} WebSocket connection closed in receive loop: {e}")
                 break
             except Exception as e:
                 logger.error(f"{self} Unexpected error in receive loop: {e}")

src/pipecat/services/mcp_service.py

@@ -7,7 +7,7 @@
 """MCP (Model Context Protocol) client for integrating external tools with LLMs."""
 
 import json
-from typing import Any, Dict, List, Tuple
+from typing import Any, Dict, List, TypeAlias
 
 from loguru import logger
 
@@ -28,6 +28,8 @@ except ModuleNotFoundError as e:
     logger.error("In order to use an MCP client, you need to `pip install pipecat-ai[mcp]`.")
     raise Exception(f"Missing module: {e}")
 
+ServerParameters: TypeAlias = StdioServerParameters | SseServerParameters | StreamableHttpParameters
+
 
 class MCPClient(BaseObject):
     """Client for Model Context Protocol (MCP) servers.
@@ -42,7 +44,7 @@ class MCPClient(BaseObject):
 
     def __init__(
         self,
-        server_params: Tuple[StdioServerParameters, SseServerParameters, StreamableHttpParameters],
+        server_params: ServerParameters,
         **kwargs,
     ):
         """Initialize the MCP client with server parameters.

src/pipecat/transports/smallwebrtc/request_handler.py

@@ -0,0 +1,200 @@
+#
+# Copyright (c) 2024–2025, Daily
+#
+# SPDX-License-Identifier: BSD 2-Clause License
+#
+
+"""SmallWebRTC request handler for managing peer connections.
+
+This module provides a client for handling web requests and managing WebRTC connections.
+"""
+
+import asyncio
+from dataclasses import dataclass
+from enum import Enum
+from typing import Any, Awaitable, Callable, Dict, List, Optional
+
+from fastapi import HTTPException
+from loguru import logger
+
+from pipecat.transports.smallwebrtc.connection import IceServer, SmallWebRTCConnection
+
+
+@dataclass
+class SmallWebRTCRequest:
+    """Small WebRTC transport session arguments for the runner.
+
+    Parameters:
+        sdp: The SDP string (Session Description Protocol).
+        type: The type of the SDP, either "offer" or "answer".
+        pc_id: Optional identifier for the peer connection.
+        restart_pc: Optional whether to restart the peer connection.
+        request_data: Optional custom data sent by the customer.
+    """
+
+    sdp: str
+    type: str
+    pc_id: Optional[str] = None
+    restart_pc: Optional[bool] = None
+    request_data: Optional[Any] = None
+
+
+class ConnectionMode(Enum):
+    """Enum defining the connection handling modes."""
+
+    SINGLE = "single"  # Only one active connection allowed
+    MULTIPLE = "multiple"  # Multiple simultaneous connections allowed
+
+
+class SmallWebRTCRequestHandler:
+    """SmallWebRTC request handler for managing peer connections.
+
+    This class is responsible for:
+      - Handling incoming SmallWebRTC requests.
+      - Creating and managing WebRTC peer connections.
+      - Supporting ESP32-specific SDP munging if enabled.
+      - Invoking callbacks for newly initialized connections.
+      - Supporting both single and multiple connection modes.
+    """
+
+    def __init__(
+        self,
+        ice_servers: Optional[List[IceServer]] = None,
+        esp32_mode: bool = False,
+        host: Optional[str] = None,
+        connection_mode: ConnectionMode = ConnectionMode.MULTIPLE,
+    ) -> None:
+        """Initialize a SmallWebRTC request handler.
+
+        Args:
+            ice_servers (Optional[List[IceServer]]): List of ICE servers to use for WebRTC
+                connections.
+            esp32_mode (bool): If True, enables ESP32-specific SDP munging.
+            host (Optional[str]): Host address used for SDP munging in ESP32 mode.
+                Ignored if `esp32_mode` is False.
+            connection_mode (ConnectionMode): Mode of operation for handling connections.
+                SINGLE allows only one active connection, MULTIPLE allows several.
+        """
+        self._ice_servers = ice_servers
+        self._esp32_mode = esp32_mode
+        self._host = host
+        self._connection_mode = connection_mode
+
+        # Store connections by pc_id
+        self._pcs_map: Dict[str, SmallWebRTCConnection] = {}
+
+    def _check_single_connection_constraints(self, pc_id: Optional[str]) -> None:
+        """Check if the connection request satisfies single connection mode constraints.
+
+        Args:
+            pc_id: The peer connection ID from the request
+
+        Raises:
+            HTTPException: If constraints are violated in single connection mode
+        """
+        if self._connection_mode != ConnectionMode.SINGLE:
+            return
+
+        if not self._pcs_map:  # No existing connections
+            return
+
+        # Get the existing connection (should be only one in single mode)
+        existing_connection = next(iter(self._pcs_map.values()))
+
+        if existing_connection.pc_id != pc_id and pc_id:
+            logger.warning(
+                f"Connection pc_id mismatch: existing={existing_connection.pc_id}, received={pc_id}"
+            )
+            raise HTTPException(status_code=400, detail="PC ID mismatch with existing connection")
+
+        if not pc_id:
+            logger.warning(
+                "Cannot create new connection: existing connection found but no pc_id received"
+            )
+            raise HTTPException(
+                status_code=400,
+                detail="Cannot create new connection with existing connection active",
+            )
+
+    async def handle_web_request(
+        self,
+        request: SmallWebRTCRequest,
+        webrtc_connection_callback: Callable[[Any], Awaitable[None]],
+    ) -> None:
+        """Handle a SmallWebRTC request and resolve the pending answer.
+
+        This method will:
+          - Reuse an existing WebRTC connection if `pc_id` exists.
+          - Otherwise, create a new `SmallWebRTCConnection`.
+          - Invoke the provided callback with the connection.
+          - Manage ESP32-specific munging if enabled.
+          - Enforce single/multiple connection mode constraints.
+
+        Args:
+            request (SmallWebRTCRequest): The incoming WebRTC request, containing
+                SDP, type, and optionally a `pc_id`.
+            webrtc_connection_callback (Callable[[Any], Awaitable[None]]): An
+                asynchronous callback function that is invoked with the WebRTC connection.
+
+        Raises:
+            HTTPException: If connection mode constraints are violated
+            Exception: Any exception raised during request handling or callback execution
+                will be logged and propagated.
+        """
+        try:
+            pc_id = request.pc_id
+
+            # Check connection mode constraints first
+            self._check_single_connection_constraints(pc_id)
+
+            # After constraints are satisfied, get the existing connection if any
+            existing_connection = self._pcs_map.get(pc_id) if pc_id else None
+
+            if existing_connection:
+                pipecat_connection = existing_connection
+                logger.info(f"Reusing existing connection for pc_id: {pc_id}")
+                await pipecat_connection.renegotiate(
+                    sdp=request.sdp,
+                    type=request.type,
+                    restart_pc=request.restart_pc or False,
+                )
+            else:
+                pipecat_connection = SmallWebRTCConnection(ice_servers=self._ice_servers)
+                await pipecat_connection.initialize(sdp=request.sdp, type=request.type)
+
+                @pipecat_connection.event_handler("closed")
+                async def handle_disconnected(webrtc_connection: SmallWebRTCConnection):
+                    logger.info(f"Discarding peer connection for pc_id: {webrtc_connection.pc_id}")
+                    self._pcs_map.pop(webrtc_connection.pc_id, None)
+
+                # Invoke callback provided in runner arguments
+                try:
+                    await webrtc_connection_callback(pipecat_connection)
+                    logger.debug(
+                        f"webrtc_connection_callback executed successfully for peer: {pipecat_connection.pc_id}"
+                    )
+                except Exception as callback_error:
+                    logger.error(
+                        f"webrtc_connection_callback failed for peer {pipecat_connection.pc_id}: {callback_error}"
+                    )
+
+            answer = pipecat_connection.get_answer()
+
+            if self._esp32_mode and self._host and self._host != "localhost":
+                from pipecat.runner.utils import smallwebrtc_sdp_munging
+
+                answer["sdp"] = smallwebrtc_sdp_munging(answer["sdp"], self._host)
+
+            self._pcs_map[answer["pc_id"]] = pipecat_connection
+
+            return answer
+        except Exception as e:
+            logger.error(f"Error processing SmallWebRTC request: {e}")
+            logger.debug(f"SmallWebRTC request details: {request}")
+            raise
+
+    async def close(self):
+        """Clear the connection map."""
+        coros = [pc.disconnect() for pc in self._pcs_map.values()]
+        await asyncio.gather(*coros)
+        self._pcs_map.clear()