Paul Kompfner
0dac7f7e48
Update OpenAIRealtimeLLMService to work with LLMContext and LLMContextAggregatorPair
2025-10-01 10:18:07 -04:00
Paul Kompfner
a2c69fbd8b
Update OpenAIRealtimeLLMService to work with LLMContext and LLMContextAggregatorPair
2025-10-01 09:57:51 -04:00
Paul Kompfner
8962263329
Update OpenAIRealtimeLLMService to work with LLMContext and LLMContextAggregatorPair
2025-09-30 14:47:06 -04:00
Paul Kompfner
236ac93ac6
Remove remaining usage of OpenAILLMContext throughout the codebase in favor of LLMContext, except for:
...
- Usage in classes that are already deprecated
- Usage related to realtime LLMs, which don't yet support `LLMContext`
- Usage in (soon-to-be-deprecated) code paths related to `OpenAILLMContext` itself and associated machinery
2025-09-30 14:33:05 -04:00
Paul Kompfner
ceba27e696
Address some AWSNovaSonicLLMService context-recording edge cases
2025-09-29 14:51:09 -04:00
Paul Kompfner
07ba02a491
Get rid of LLMContext.get_messages_for_persistent_storage().
...
The reason for its `system_instruction` argument was to support usage with LLMs where you might pass the system instruction as a parameter to the `LLMService` rather than specifying it in the context.
But as I thought about it more I became unconvinced that the `system_instruction` argument was really beneficial:
- If you specified your system instruction in your context in the first place, it'll still be there when you read messages for persistent storage
- If you didn't specify your system instruction in the context and instead passed it in as an `LLMService` parameter, you most likely *don't* want it to be in the context when you read messages for persistent storage
- ...and if you really really do need to inject it at the start of the context, it's quite easy to do anyway
And if we remove the `system_instruction` argument from `get_messages_for_persistent_storage()`, then it's essentially just `get_messages()`.
2025-09-29 11:00:02 -04:00
Paul Kompfner
4d9873b613
Update AWSNovaSonicLLMService to work with LLMContext and LLMContextAggregatorPair
2025-09-29 10:28:38 -04:00
Aleix Conchillo Flaqué
f53fd880dc
BaseInputTransport: create VAD thread in VADAnalyzer
...
We move the thread creation to the VADAnalyzer instead of the input
transport. This can potentially be useful if we need to analyze multiple audio
streams.
2025-09-25 13:41:20 -07:00
Aleix Conchillo Flaqué
7bafc3a1bb
BaseSmartTurn: process speech in a separate thread
2025-09-25 13:37:28 -07:00
Aleix Conchillo Flaqué
22ef61fe8d
BaseTurnAnalyzer: add BaseTurnParams base class for parameters
2025-09-25 13:37:09 -07:00
Aleix Conchillo Flaqué
33447ad6f2
BaseOpenAILLMService: include cached tokens to metrics frame
2025-09-24 19:32:16 -07:00
Aleix Conchillo Flaqué
cb048d6c7e
Merge pull request #2733 from pipecat-ai/aleix/tavus-missing-daily-callback
...
TavusTransport: add missing on_before_leave callback
2025-09-24 11:10:32 -07:00
Aleix Conchillo Flaqué
6c2c43ade0
Merge pull request #2724 from pipecat-ai/pk/update-natural-conversation-examples-with-universal-context
...
Update natural conversation examples with universal context
2025-09-24 11:07:50 -07:00
Aleix Conchillo Flaqué
f899c15b03
Merge pull request #2731 from pipecat-ai/pk/update-example-25-to-use-universal-context
...
Update example 25 to use universal `LLMContext`
2025-09-24 11:05:36 -07:00
Aleix Conchillo Flaqué
d10ef08775
Merge pull request #2727 from pipecat-ai/pk/strands-agents-needs-to-support-openaillmcontextframe-for-now
...
`StrandsAgentsProcessor` should still support `OpenAILLMContextFrame`…
2025-09-24 11:05:07 -07:00
Aleix Conchillo Flaqué
4bff0a7c49
Merge pull request #2732 from pipecat-ai/pk/update-voicemail-detector-to-use-llm-context
...
Update `VoicemailDetector` to use universal `LLMContext`
2025-09-24 11:04:42 -07:00
Aleix Conchillo Flaqué
508f7d203d
Merge pull request #2729 from pipecat-ai/aleix/frame-processor-cancel-default-timeout
...
FrameProcessor: timeout when cancelling tasks
2025-09-24 10:55:52 -07:00
Filipi Fuchter
f6164e3bde
HeyGen implementation for Pipecat - HeyGenTransport.
2025-09-24 14:33:56 -03:00
Aleix Conchillo Flaqué
1a0fb55d0f
TavusTransport: add missing on_before_leave callback
2025-09-24 10:18:56 -07:00
Paul Kompfner
6d0beef944
Update VoicemailDetector to use universal LLMContext
2025-09-24 12:58:42 -04:00
Paul Kompfner
dea0f1791f
Update OpenAILLMAdapter.get_messages_for_logging() to truncate "input_audio" message data
2025-09-24 12:41:41 -04:00
Aleix Conchillo Flaqué
f9eb447d82
FrameProcessor: timeout when cancelling tasks
2025-09-24 08:24:28 -07:00
Paul Kompfner
65f5fe8588
StrandsAgentsProcessor should still support OpenAILLMContextFrame until that frame has been deprecated
2025-09-24 11:05:27 -04:00
Paul Kompfner
817c77f3fe
Update SmallWebRTCTransport to pass a sender ID in the "on_app_message" event
2025-09-24 10:24:59 -04:00
Paul Kompfner
66b7977a62
Make SmallWebRTCTransport adhere to the expected "on_app_message" event signature
2025-09-24 09:34:28 -04:00
Mark Backman
468de68aec
Fixup for TwilioFrameSerializer
2025-09-24 09:32:46 -04:00
Jessie Wei
88614b312f
Merge branch 'pipecat-ai:main' into main
2025-09-24 10:23:52 +10:00
Jessie Wei
5b4655f45a
chore: Update per review comments
2025-09-24 00:22:56 +00:00
Aleix Conchillo Flaqué
d7c8f8df53
update CHANGELOG with AudioBufferProcessor fixes
2025-09-23 15:42:01 -07:00
Aleix Conchillo Flaqué
15782be27c
Merge pull request #2676 from golbin/main
...
Fix audio buffer flush and silence handling
2025-09-23 15:27:31 -07:00
Aleix Conchillo Flaqué
620b1f785c
examples: update Strands Agents with universal context and add evals
2025-09-23 11:37:57 -07:00
Aleix Conchillo Flaqué
17ea0afa6f
StrandsAgentsProcessor: more formatting fixes
2025-09-23 11:05:14 -07:00
Aleix Conchillo Flaqué
3fc5214c15
BaseOutputTransport: only push downstream if transport write successful
...
Fixes #2589
2025-09-23 11:04:37 -07:00
Aleix Conchillo Flaqué
ed64716219
StrandsAgentsProcessor: fix formatting
2025-09-23 10:38:31 -07:00
Aleix Conchillo Flaqué
5c22b2e1de
Merge pull request #2610 from adithyaxx/add-strands-processor
...
Add native Strands Agents support to Pipecat
2025-09-23 10:33:22 -07:00
Mark Backman
40c82a8530
Merge pull request #2716 from pipecat-ai/mb/add-11labs-stt
...
Add ElevenLabsSTTService
2025-09-23 12:21:08 -04:00
kompfner
88337fc21f
Merge pull request #2717 from pipecat-ai/pk/mem0-support-univeral-context
...
Add support for universal `LLMContext` to `Mem0MemoryService`
2025-09-23 11:54:55 -04:00
kompfner
928c0ef1b4
Merge pull request #2715 from pipecat-ai/pk/langchain-processor-support-univeral-context
...
Add support for universal `LLMContext` to `LangchainProcessor`
2025-09-23 11:54:44 -04:00
kompfner
1f005e7075
Merge pull request #2714 from pipecat-ai/pk/gated-openai-llm-context-aggregator-support-univeral-context
...
Add support for universal `LLMContext` to `GatedOpenAILLMContextAggre…
2025-09-23 11:54:26 -04:00
kompfner
2cee6229ae
Merge pull request #2713 from pipecat-ai/pk/log-observer-support-univeral-context
...
Add support for universal `LLMContext` to `LLMLogObserver`
2025-09-23 11:54:11 -04:00
Aleix Conchillo Flaqué
10e9371f49
Merge pull request #2712 from pipecat-ai/aleix/pipeline-runner-signals-windows
...
PipelineRunner: use signal.signal() on Windows
2025-09-23 08:28:26 -07:00
Paul Kompfner
e21ab89509
Add support for universal LLMContext to Mem0MemoryService
2025-09-23 10:35:54 -04:00
Mark Backman
cbce2075eb
Add ElevenLabsSTTService
2025-09-23 10:27:31 -04:00
Paul Kompfner
97868175e6
Add support for universal LLMContext to LangchainProcessor
2025-09-23 10:00:48 -04:00
Paul Kompfner
99731ca40a
Add support for universal LLMContext to GatedOpenAILLMContextAggregator, renaming it to GatedLLMContextAggregator in the process
2025-09-23 09:52:13 -04:00
Paul Kompfner
f96cbcce22
Add support for universal LLMContext to LLMLogObserver
2025-09-23 09:30:14 -04:00
Filipi da Silva Fuchter
9b4f1bdf39
Merge pull request #2705 from tzookb/tzookb/latency-logging
...
update UserBotLatencyLogObserver to have logging in functions that can be overidden
2025-09-23 09:46:20 -03:00
Filipi Fuchter
6b2bf8de64
Fixing the ruff format and making the methods sync.
2025-09-23 09:41:50 -03:00
Filipi Fuchter
a5776b20ad
Monitoring the peer connection while it is in the *connecting* state.
2025-09-23 08:30:17 -03:00
Filipi Fuchter
eaecefe675
Refactoring how we are reading the image bytes inside the base llm.
2025-09-23 08:05:15 -03:00
