wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
7,500 Commits 276 Branches 111 Tags
fa85f7bbc73d2ad7570f7516d44e59bfb6c6b1ac
Commit Graph

4335 Commits

Author SHA1 Message Date
Mark Backman
fa85f7bbc7 Merge pull request #3640 from lukepayyapilli/fix/openai-stream-close 
fix: close stream on cancellation to prevent socket leaks
 2026-02-05 18:00:06 -05:00
filipi87
76f63e54e2 Changing the ‘no audio received’ log from warning to debug. 2026-02-05 18:07:14 -03:00
Filipi da Silva Fuchter
cce5a13444 Merge pull request #3650 from pipecat-ai/filipi/twilio_issues 
Ignoring RTVI messages inside the Serializers by default.
 2026-02-05 15:52:59 -05:00
filipi87
6af4d872a8 Refactoring the serializers to ignore the RTVI messages by default. 2026-02-05 16:52:53 -03:00
Mark Backman
56d8ef2bf4 Deprecate UserBotLatencyLogObserver, update 29 example 2026-02-05 14:29:45 -05:00
Derek Haynes
93138466d6 Feat: Add user-bot latency to OTel turn spans 
This adds user-to-bot response latency tracking to OpenTelemetry spans:

- Created UserBotLatencyObserver as a reusable component for tracking
user-to-bot response latency
- Records the value as an attribute on turn spans (turn.user_bot_latency_seconds)
- Updated TurnTraceObserver to use UserBotLatencyObserver, following the same pattern as TurnTrackingObserver
- Updated PipelineTask to automatically create and wire UserBotLatencyObserver
when tracing is enabled (same as TurnTrackingObserver)
 2026-02-05 14:29:42 -05:00
Mark Backman
5a5a98b497 Merge pull request #3649 from itsderek23/fix/tracing-orphan-spans 
Fix orphan otel spans during flow initialization and transitions
 2026-02-05 14:23:52 -05:00
Derek Haynes
0d45b48f7b Fix import placement 2026-02-05 11:26:58 -07:00
Mark Backman
bd12b60b5c Merge pull request #3614 from okue/fix/websocket-broadcast-frame-misuse 
fix: pass frame class instead of instance to broadcast_frame in websocket transports
 2026-02-05 12:19:03 -05:00
filipi87
752e16f553 Ignoring RTVI messages inside TwilioSerializer by default. 2026-02-05 10:51:03 -03:00
Derek Haynes
7c7408a048 Fix orphan spans in tracing during flow initialization and transitions 
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-02-05 06:06:13 -07:00
Mark Backman
8f42343927 Merge pull request #3630 from pipecat-ai/mb/add-function-call-messages-rtvi 
Add native RTVI function call lifecycle messages
 2026-02-04 16:20:42 -05:00
Mark Backman
ecb02d9049 Bump RTVI_PROTOCOL_VERSION to 1.2.0 2026-02-04 11:17:38 -05:00
Mark Backman
cc68e00125 Deprecate llm-function-call message 2026-02-04 11:17:23 -05:00
Mark Backman
e0e3b5250b Add RTVIObserverParams to control what information is included in function call events 2026-02-04 11:05:05 -05:00
Luke Payyapilli
55a3b10e70 fix(openai): close stream on cancellation to prevent socket leaks 2026-02-04 09:59:10 -05:00
Mark Backman
2db3d94d06 Merge pull request #3628 from pipecat-ai/mb/broadcast-speech-control-params-frame 
Fix: Broadcast SpeechControlParamsFrame from VADController
 2026-02-03 18:44:15 -05:00
Mark Backman
2a26b9f7a3 Fix: Broadcast SpeechControlParamsFrame from VADController 2026-02-03 18:40:39 -05:00
Aleix Conchillo Flaqué
c3a4da4a29 PipelineTask: also call set_bot_ready() for external RTVI processors 2026-02-03 14:16:08 -08:00
Mark Backman
84ca0b6d58 Merge pull request #3629 from pipecat-ai/fix/telephony-websocket-stopasynciteration 
Fix StopAsyncIteration in parse_telephony_websocket
 2026-02-03 12:10:07 -05:00
Mark Backman
c1857d255d Avoid nesting try/excepts 2026-02-03 12:00:04 -05:00
Mark Backman
40c84faff5 Remove handle_function_call_start 2026-02-03 10:00:59 -05:00
Mark Backman
84cd9346f9 Add native RTVI function call lifecycle messages 2026-02-03 10:00:59 -05:00
Luke Payyapilli
8d3e10f054 Make EndFrame and StopFrame uninterruptible to prevent pipeline freeze 2026-02-03 09:12:59 -05:00
James Hush
803a20cc00 Fix formatting: remove extra blank line 
Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
 2026-02-03 16:46:44 +08:00
James Hush
90bead06ab Apply suggestions from code review 
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
 2026-02-03 16:42:13 +08:00
James Hush
b030f1178d Add changelog and improve docstring for parse_telephony_websocket 
- Added changelog entry for bug fix
- Enhanced docstring with Args and Raises sections

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
 2026-02-03 16:26:09 +08:00
James Hush
a627597bca Fix StopAsyncIteration in parse_telephony_websocket 
Handle WebSocket disconnections gracefully when telephony providers send
fewer messages than expected. Adds explicit StopAsyncIteration handling
for both first and second message retrieval.

Co-Authored-By: Claude Sonnet 4.5 <noreply@anthropic.com>
 2026-02-03 16:25:07 +08:00
Mark Backman
a4e499dc80 Merge pull request #3617 from pipecat-ai/fix/cjk-sentence-splitting 
Fix sentence splitting for CJK and other non-Latin languages
 2026-02-02 18:16:51 -05:00
Mark Backman
a592b7fdf0 Update per PR 1789, align with ErrorFrame norms 2026-02-02 08:55:29 -05:00
Mark Backman
ba2b7c05d6 Add ResembleAITTSService 2026-02-02 08:55:27 -05:00
James Hush
763002f2bc Fix sentence splitting for CJK and other non-Latin languages in TTS pipeline 
NLTK's sent_tokenize() only supports ~15 European languages and defaults to
English. For Japanese, Chinese, Korean, Hindi, Arabic, and other non-Latin
languages, NLTK fails to recognize sentence boundaries like 。？！ causing
text to accumulate until flush instead of being emitted sentence-by-sentence.

Add a fallback in match_endofsentence() that scans for unambiguous non-Latin
sentence-ending punctuation when NLTK fails to split the text. Latin
punctuation (. ! ? ; …) is excluded from the fallback since NLTK handles
those correctly and they can be ambiguous (abbreviations, decimals, etc.).

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
 2026-02-02 14:27:49 +08:00
okue
d3ecbb11c1 fix: pass frame class instead of instance to broadcast_frame in websocket transports 
broadcast_frame() expects a frame class and kwargs, but the three
websocket input transports (fastapi, client, server) were incorrectly
passing a frame instance. This would cause a TypeError at runtime when
an InputTransportMessageFrame was received.
 2026-02-01 20:38:34 +09:00
Aleix Conchillo Flaqué
f453227ba3 Merge pull request #3612 from pipecat-ai/aleix/use-kokoro-onnx 
KokoroTTSService: use kokoro-onnx instead of kokoro
 2026-01-31 21:03:55 -08:00
Aleix Conchillo Flaqué
95689cc81c KokoroTTSService: use kokoro-onnx instead of kokoro 2026-01-31 17:20:27 -08:00
Aleix Conchillo Flaqué
acc9923c0a PipelineTask: don't add RTVIObserver if already there 2026-01-31 14:54:29 -08:00
Aleix Conchillo Flaqué
d87f3543c7 GeminiLiveLLMService: let the transcription timeout handler be scheduled 2026-01-30 22:41:10 -08:00
Mark Backman
e779233918 Fix IVRNavigator to push AggregatedTextFrame when switching to conversation mode 2026-01-30 21:07:49 -05:00
Aleix Conchillo Flaqué
e3158e1131 Merge pull request #3600 from pipecat-ai/aleix/llm-server-timeout-task-never-waited 
LLMService: make sure function call timeout handler is started
 2026-01-30 12:01:18 -08:00
Mark Backman
63a23246d5 Add UserTurnCompletionLLMServiceMixin (#3518) 
* Added UserTurnCompletionLLMServiceMixin class

* Added 22-filter-incomplete-turns.py foundational example

* Removed old 22 natural conversation foundational examples

* Added test_user_turn_completion_mixin.py
 2026-01-30 14:57:15 -05:00
Aleix Conchillo Flaqué
a98ca9b65b LLMService: make sure function call timeout handler is started 2026-01-30 11:38:26 -08:00
Aleix Conchillo Flaqué
35d265770d LLMUserAggregator: don't process certain self-queued frames 2026-01-30 10:07:34 -08:00
Aleix Conchillo Flaqué
9632efec8c VADProcessor: broadcast frames 2026-01-30 10:07:34 -08:00
Aleix Conchillo Flaqué
27dbfa1eda NvidiaTTSService: return AsyncIterator instead of AsyncIterable 2026-01-30 10:07:34 -08:00
Aleix Conchillo Flaqué
183c0aa4ef LLMUserAggregator: queue frames internally so strategies and controllers can process them 2026-01-30 10:07:34 -08:00
Aleix Conchillo Flaqué
c46e7f5da0 TurnAnalyzerUserTurnStopStrategy: only update vad params if frame contains vad 2026-01-30 10:07:34 -08:00
Aleix Conchillo Flaqué
b486f35c70 audio: add new VADProcessor 2026-01-30 10:07:34 -08:00
Aleix Conchillo Flaqué
c92080b0d2 LLMUserAggregator: add vad_analyzer and use VADController 2026-01-30 10:07:34 -08:00
Aleix Conchillo Flaqué
ddfedaf478 audio(vad): add new VADController 2026-01-30 10:07:34 -08:00
Aleix Conchillo Flaqué
b1ad4d5ab0 BaseInputTransport: deprecate vad_analyzer 2026-01-30 10:07:33 -08:00