Mark Backman
82f0896d6a
Propagate Soniox token language
2026-05-13 15:23:22 -04:00
Mark Backman
08680732f6
Merge pull request #4475 from pipecat-ai/mb/cartesia-korean-fix
...
Fix Cartesia CJK timestamp spacing
2026-05-13 13:20:42 -04:00
Mark Backman
064b68aa01
Fix Cartesia CJK timestamp spacing
2026-05-13 13:13:40 -04:00
Filipi da Silva Fuchter
b0f8ea7e28
Merge pull request #4477 from pipecat-ai/filipi/nvidia_sagemaker_follow_up
...
NVidia TTS Sagemaker: Buffering audio to avoid glitches.
2026-05-13 14:06:44 -03:00
filipi87
ad50c8d5d5
Buffering audio to avoid glitches.
2026-05-13 14:01:03 -03:00
Mark Backman
5fef239b68
Merge pull request #4450 from pipecat-ai/mb/gpt-realtime-whisper
...
Default OpenAI Realtime transcription to gpt-realtime-whisper
2026-05-13 09:48:33 -04:00
Filipi da Silva Fuchter
9148e307cc
Merge pull request #4464 from pipecat-ai/filipi/nvidia_sagemaker
...
NVidia sagemaker - TTS and STT services
2026-05-13 07:53:26 -03:00
Filipi da Silva Fuchter
703d23b658
Update examples/voice/voice-nvidia-sagemaker.py
...
Co-authored-by: Mark Backman <mark@daily.co >
2026-05-13 06:36:57 -04:00
Filipi da Silva Fuchter
227ba288da
Update examples/voice/voice-nvidia-sagemaker.py
...
Co-authored-by: Mark Backman <mark@daily.co >
2026-05-13 06:36:45 -04:00
Mark Backman
3e8c5c08f4
Clarify realtime settings update condition
2026-05-12 17:48:53 -04:00
Mark Backman
644030584f
Centralize OpenAI audio constants
2026-05-12 17:48:53 -04:00
filipi87
0740021ff4
Removing changelog for sanitize_text_for_tts
2026-05-12 18:29:35 -03:00
filipi87
68f265fa62
Fixing ruff format.
2026-05-12 18:28:14 -03:00
filipi87
b9f052079d
Removing sanitize_text_for_tts
2026-05-12 18:22:15 -03:00
filipi87
130bb7371c
Removing sanitize_text_for_tts
2026-05-12 18:21:47 -03:00
filipi87
5d61763987
Refactoring how we are reconnecting the STT.
2026-05-12 18:20:19 -03:00
filipi87
7984556692
Fixing typecheck.
2026-05-12 18:00:07 -03:00
filipi87
bea9e4b3ba
New example voice-nvidia-sagemaker.py
2026-05-12 17:44:11 -03:00
Mark Backman
19df443500
Merge pull request #4471 from pipecat-ai/mb/fix-gstreamer-pyright-import
2026-05-12 16:34:48 -04:00
Mark Backman
07f241143b
Merge pull request #4469 from pipecat-ai/mb/remove-vad-analyzer-runner-utils-docstring
2026-05-12 16:34:27 -04:00
Mark Backman
2fdb9bbf42
Merge pull request #4462 from pipecat-ai/mb/cartesia-sonic-3.5
2026-05-12 16:34:04 -04:00
filipi87
0146947b68
Addressing the comments left in the PR review.
2026-05-12 17:12:19 -03:00
Mark Backman
e2bfa6352f
Add changelog for #4450
2026-05-12 15:20:57 -04:00
Mark Backman
abd28e2ac1
Update OpenAI realtime transcription default
2026-05-12 15:20:57 -04:00
kompfner
88deebbf5f
Merge pull request #4472 from pipecat-ai/pk/default-gpt-realtime-2
...
Switch OpenAIRealtimeLLMService default model to gpt-realtime-2
2026-05-12 15:17:12 -04:00
filipi87
c2bdc1aada
Fixing metrics and adding extra guard after sanitization.
2026-05-12 16:11:01 -03:00
Paul Kompfner
fc0589e8f1
Switch OpenAIRealtimeLLMService default model to gpt-realtime-2
2026-05-12 14:57:59 -04:00
kompfner
67f8d34e9f
Merge pull request #4470 from pipecat-ai/pk/gpt-realtime-2-reasoning-effort
...
Add reasoning support to OpenAIRealtimeLLMService for gpt-realtime-2
2026-05-12 14:43:39 -04:00
kompfner
d3b8710720
Merge pull request #4465 from pipecat-ai/pk/gpt-realtime-2
...
Handle gpt-realtime-2 multi-output-item audio responses
2026-05-12 14:30:15 -04:00
Mark Backman
86e2aa85d3
Fix GStreamer pipeline source pyright import
2026-05-12 14:16:36 -04:00
Paul Kompfner
b89500256d
Drop debug logging added while investigating multi-output-item audio
2026-05-12 14:05:16 -04:00
Paul Kompfner
a52bdef32b
Add reasoning support to OpenAIRealtimeLLMService for gpt-realtime-2
2026-05-12 13:55:19 -04:00
Mark Backman
afd9fc5fdf
Remove vad_analyzer from create_transport docstring example
2026-05-12 13:50:17 -04:00
filipi87
7f98dba925
Changelog files for the new nvidia features.
2026-05-12 14:43:12 -03:00
filipi87
6a27ed35b1
Fixing the Bidi client to accept None.
2026-05-12 12:19:30 -03:00
filipi87
a34864d643
Fixed ruff, pyright, and test_service_init failures
2026-05-12 11:39:52 -03:00
Paul Kompfner
007fa3a3a8
Handle gpt-realtime-2 multi-output-item audio responses
...
A single Realtime API response can now contain more than one audio item
(observed with gpt-realtime-2), and the first item's audio.done can
arrive after deltas from the second have started arriving. Deltas still
arrive strictly in playback order across items, so we keep forwarding
them as received — matching OpenAI's reference implementation.
Adjusted OpenAIRealtimeLLMService so a multi-item response is treated as
one continuous TTS turn:
- _handle_evt_audio_delta: on item switch, advance the tracked item in
place (reset total_size) without emitting another TTSStartedFrame.
Truncation now always targets the latest item.
- _handle_evt_audio_done: debug-trace only; no longer pushes
TTSStoppedFrame.
- _handle_evt_response_done: pushes a single TTSStoppedFrame per turn,
bookending the audio with the Started pushed on the first delta.
Added tests covering single-item, overlapping multi-item, non-overlapping
multi-item, and interrupt-during-multi-item (last-item-wins truncation).
2026-05-12 10:34:50 -04:00
filipi87
5dd7413c00
Nvidia Sagemaker Nemotron ASR STT service
2026-05-12 11:16:00 -03:00
filipi87
8e0a338d96
Nvidia Sagemaker Magpie TTS service
2026-05-12 11:15:42 -03:00
Mark Backman
d65aee9181
Add changelog for #4462
2026-05-11 17:34:00 -04:00
Mark Backman
1755016679
Update default Cartesia TTS model to sonic-3.5
2026-05-11 17:33:40 -04:00
Mark Backman
b7f6298601
Merge pull request #4461 from pipecat-ai/mb/security-vuln-2025-05-11
...
Update uv.lock for urllib3 and langchain-core
2026-05-11 15:58:05 -04:00
Mark Backman
396873ac7e
Merge pull request #4460 from pipecat-ai/mb/codex-skills
...
Add Codex skills and AGENTS.md
2026-05-11 15:57:49 -04:00
Mark Backman
5b33964a1b
Update uv.lock for urllib3 and langchain-core
2026-05-11 15:51:01 -04:00
Mark Backman
8b37cd1d3a
Add agent-neutral repository instructions
2026-05-11 15:43:43 -04:00
Mark Backman
7a2b667fa1
Add Codex skill symlinks
2026-05-11 15:27:49 -04:00
Mark Backman
ee8c607315
Merge pull request #4452 from pipecat-ai/mb/cleanup-frontmatter
...
Add cleanup skill frontmatter
2026-05-11 09:33:44 -04:00
Aleix Conchillo Flaqué
71578e7151
Merge pull request #4449 from pipecat-ai/aleix/base-object-task-manager
...
Move create_task and cancel_task from FrameProcessor to BaseObject
2026-05-10 20:36:54 -07:00
Aleix Conchillo Flaqué
77058b01c4
Add changelog for #4449
2026-05-10 20:34:52 -07:00
Aleix Conchillo Flaqué
4f85e7c089
Fix pyright cr_code access on Coroutine in BaseObject.create_task
...
`collections.abc.Coroutine` doesn't expose `cr_code`/`co_name`; only
native coroutine objects do. Use `getattr` chains so pyright is happy
and any non-native awaitable falls back to a generic task name instead
of crashing.
2026-05-10 20:34:52 -07:00
