wx44wx/pipecat
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
7,680 Commits 276 Branches 111 Tags
1daea78b913fe2ca4f64efc1038eaedcfefba9e7
Commit Graph

347 Commits

Author SHA1 Message Date
Mark Backman
f181e12d8f Add changelog for PR #3759 2026-02-16 11:35:45 -07:00
Mark Backman
dba4de77bf Merge pull request #3684 from ai-coustics/goedev/aic-model-caching 
AIC model caching
 2026-02-16 10:43:14 -05:00
Mark Backman
c682a44bb6 Merge pull request #3738 from lukepayyapilli/fix/mute-events-before-start-frame 
Fix LLMUserAggregator broadcasting mute events before StartFrame
 2026-02-14 09:40:40 -05:00
Mark Backman
79ec61d1d8 Merge pull request #3642 from pipecat-ai/cb/rime-arcana-v3 
Update RimeTTSService for arcana and mistv2 model support
 2026-02-13 13:25:27 -05:00
Mark Backman
ca440594fe Merge pull request #3720 from pipecat-ai/mb/fix-grok-realtime 
Fix Grok Realtime voice type validation for server responses
 2026-02-13 13:24:53 -05:00
Mark Backman
09bb6bb03b Merge pull request #3735 from pipecat-ai/mb/fix-llm-tracing-error-handilng 
Fix double execution of service functions on tracing errors
 2026-02-13 13:23:55 -05:00
Mark Backman
746fdfbfef Merge pull request #3728 from pipecat-ai/mb/upgrade-pillow 
Bump Pillow upper bound from <12 to <13
 2026-02-13 13:23:41 -05:00
Mark Backman
a5f95acaf5 Add changelog for PR #3735 2026-02-13 13:08:03 -05:00
Luke Payyapilli
3adb2f50a6 Fix LLMUserAggregator broadcasting mute events before StartFrame 2026-02-13 11:59:56 -05:00
Mark Backman
01b7a93e08 Deprecate unused Traceable/class_decorators module and fix stale comments 
The class_decorators.py module (Traceable, @traceable, @traced) is not
used anywhere in the codebase. Mark it deprecated and fix the misleading
comment in service_decorators.py that referenced it as if it were active.
 2026-02-13 11:25:40 -05:00
Ian Lee
94f01af545 [inworld] Allow Async delivery of timestamps info 
* speed up first audio chunk latency
 2026-02-12 17:48:58 -08:00
Filipi da Silva Fuchter
432870cc36 Merge pull request #3729 from pipecat-ai/filipi/elevenlabs_issue 
TTS services fixes.
 2026-02-12 16:31:46 -05:00
Filipi da Silva Fuchter
e065907745 Merge pull request #3718 from pipecat-ai/filipi/bot_started_speaking 
Fixing an issue in RTVI where we were sometimes receiving bot output messages before the bot started speaking.
 2026-02-12 16:31:14 -05:00
filipi87
9569625f03 Changelog entries for the TTS fixes. 2026-02-12 16:11:02 -03:00
Mark Backman
18afe37bd1 Add changelog entries for PR #3642 2026-02-12 14:09:24 -05:00
Mark Backman
08beb0264a Add changelog entries for PR #3730 2026-02-12 13:14:11 -05:00
Mark Backman
4667a3d66d Add changelog for #3728 2026-02-12 09:42:23 -05:00
Mark Backman
d99a256715 Merge pull request #3706 from ianbbqzy/ian/inworld-user-agent 
[Inworld] add User-Agent and X-Request-Id for better traceability
 2026-02-11 19:38:26 -05:00
Ian Lee
dcbcab1542 [Inworld] add User-Agent and X-Request-Id for better traceability 2026-02-11 15:47:20 -08:00
Mark Backman
a966947220 Add changelog for #3720 2026-02-11 18:04:58 -05:00
filipi87
ed7fde324e Adding changelog entry for the RTVIObserver fix. 2026-02-11 16:23:42 -03:00
filipi87
0c3e59ed61 Adding changelog entry for the SmallWebRTCTransport fix. 2026-02-11 13:07:52 -03:00
Gökmen Görgen
2036757b84 add unit tests for AICModelManager and AICFilter error handling, model loading, and processor behavior 2026-02-11 15:22:37 +01:00
aconchillo
f1e2001a4e Update changelog for version 0.0.102 2026-02-10 18:28:21 -08:00
Aleix Conchillo Flaqué
f3eb5b30a0 Add changelog for #3707 2026-02-10 18:01:29 -08:00
Mark Backman
17ab9c425f Merge pull request #3675 from pipecat-ai/mb/elevenlabs-realtime-send-silence 
Add silence-based keepalive to WebsocketSTTService
 2026-02-10 18:03:38 -05:00
Mark Backman
2f5e61ac55 Add silence-based keepalive to WebsocketSTTService 
Adds opt-in keepalive_timeout and keepalive_interval params to
WebsocketSTTService. When enabled, a background task sends silent audio
(or a service-specific protocol message) when the connection has been
idle, preventing server-side timeout disconnects.

Subclasses override _send_keepalive(silence) to wrap the silence in
their wire format. The default sends raw PCM bytes.

Enables keepalive for ElevenLabs (10s), Gladia (20s), and Soniox (1s),
replacing their per-service custom keepalive tasks.
 2026-02-10 17:58:47 -05:00
filipi87
2475697955 Changelog entries for context summarization 2026-02-10 18:59:12 -03:00
Filipi da Silva Fuchter
9c627e7292 Merge pull request #3653 from pipecat-ai/filipi/heygen_lite 
HeyGen improvements.
 2026-02-10 12:12:22 -05:00
Filipi da Silva Fuchter
ad179b0852 Merge pull request #3584 from pipecat-ai/filipi/speak_frame 
TTS services improvements.
 2026-02-10 12:11:47 -05:00
filipi87
5128089d42 Add changelog entries for PR #3653. 2026-02-10 14:02:32 -03:00
filipi87
e00b98343e Changelog entries for TTS context tracking 2026-02-10 11:37:21 -03:00
Mark Backman
2dd3e2f1e7 Merge pull request #3697 from pipecat-ai/mb/soniox-rt-4 
Update SonioxSTTService default model to stt-rt-v4
 2026-02-10 09:24:39 -05:00
Mark Backman
88e981c013 Set vad_force_turn_endpoint to False in SonioxSTTService 2026-02-10 09:16:03 -05:00
Mark Backman
7bd8dfe898 Add changelog for PR 3700 2026-02-10 08:20:03 -05:00
Mark Backman
d47d95e1f0 Update SonioxSTTService default model to stt-rt-v4 2026-02-09 23:48:08 -05:00
Mark Backman
79b9d929c5 Merge pull request #3682 from eoinoreilly30/patch-1 
Add new voice options 'marin' and 'cedar'
 2026-02-09 23:47:39 -05:00
Eoin
dfc0856d54 Added changelog entry 2026-02-10 12:31:26 +09:00
Mark Backman
7684a94c33 AssemblyAISTTService: Disable turn detection when setting vad_force_turn_endpoint to True 2026-02-09 22:20:35 -05:00
Mark Backman
981253c703 Rename RequestMetadataFrame to ServiceSwitcherRequestMetadataFrame with service targeting 
Add a `service` field so the frame targets a specific service, allowing
ServiceSwitcher.push_frame to consume it only when the targeted service
matches the active service. STTService and test mocks now push the frame
downstream after handling instead of silently consuming it.
 2026-02-09 16:48:34 -05:00
Mark Backman
6305e04569 Clean up on Sarvam STT and TTS classes 2026-02-09 15:53:05 -05:00
Mark Backman
3ff9b7b5ad Merge pull request #3687 from pipecat-ai/mb/rtvi-mute-events 
Emit RTVI events for user mute/unmute
 2026-02-09 15:18:28 -05:00
Aleix Conchillo Flaqué
ca0d2e68c3 Add changelog for #3689 2026-02-09 11:58:09 -08:00
Mark Backman
00ec6c77ea Emit RTVI events for user mute/unmute state changes 
Add UserMuteStartedFrame/UserMuteStoppedFrame and corresponding RTVI
messages so clients can observe when mute strategies activate/deactivate.
 2026-02-09 14:44:32 -05:00
Mark Backman
5e66702cf5 Improved the accuracy of the UserBotLatencyObserver and UserBotLatencyLogObserver 2026-02-09 14:12:33 -05:00
Mark Backman
34b068d657 Improve user turn stop timing by triggering timeout from VAD stop 
Refactor TranscriptionUserTurnStopStrategy and TurnAnalyzerUserTurnStopStrategy
to use VADUserStoppedSpeakingFrame as the ground truth for when speech ended,
rather than triggering timeouts from transcription frames.
 2026-02-09 14:12:33 -05:00
Mark Backman
05e2a013b3 Merge pull request #3672 from pipecat-ai/mb/rtvi-duplicate-events 
Filter RTVIObserver to downstream frames only and broadcast FunctionCallCancelFrame
 2026-02-09 12:58:28 -05:00
Mark Backman
5f64dae0cf Filter RTVIObserver to downstream frames only and broadcast FunctionCallCancelFrame 
RTVIObserver now skips upstream frames to prevent duplicate RTVI messages
when frames are broadcast in both directions. Also changed
FunctionCallCancelFrame to use broadcast_frame for consistency with
other function call frames.
 2026-02-09 12:39:25 -05:00
Gökmen Görgen
ed3ec045aa add changelog file. 2026-02-09 12:04:09 +01:00
Mark Backman
570f2d7fc0 Merge pull request #3667 from ianbbqzy/ian/fix-auto-mode-space 
[inworld] aggregate_sentence mode needs trailing space
 2026-02-07 18:22:32 -05:00