v0.2.11

Mesh-LLM/mesh-llmv0.2.11Aug 3, 2026by andimarafioti

AI Summary

This patch release adds a Chat Completions backend, text-only response support, and a new realtime web demo, alongside fixes for generation races and Docker builds.

Key Highlights

New Chat Completions LLM backend supporting OpenAI `/v1/chat/completions`.
Support for text-only and out-of-band (conversation=none) responses.
New realtime web demo app added under the `demo/` directory.
Fixes for mid-generation conversation races involving tool calls and images.

New Features

Chat Completions LLM backend
Text-only and out-of-band response support
Realtime web demo app

Full Release Notes

## What's Changed
* Bump the actions group with 2 updates by @dependabot[bot] in https://github.com/huggingface/speech-to-speech/pull/316
* Support text-only and out-of-band (conversation=none) responses by @A-Mahla in https://github.com/huggingface/speech-to-speech/pull/318
* Emit assistant transcript for fresh response when discard guard is stuck by @A-Mahla in https://github.com/huggingface/speech-to-speech/pull/321
* Add chat-completions LLM backend (OpenAI /v1/chat/completions) by @A-Mahla in https://github.com/huggingface/speech-to-speech/pull/322
* Default Qwen3 TTS to GGML by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/325
* Fix two mid-generation conversation races (tool call & image) by @A-Mahla in https://github.com/huggingface/speech-to-speech/pull/326
* Defer client conversation items during an active response by @A-Mahla in https://github.com/huggingface/speech-to-speech/pull/327
* Update HF router Gemma chat completions example by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/328
* Refresh README docs by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/330
* chore: remove deprecated melo option from tts arguments by @abhimanyus1997 in https://github.com/huggingface/speech-to-speech/pull/329
* Bump the actions group with 2 updates by @dependabot[bot] in https://github.com/huggingface/speech-to-speech/pull/324
* Fix Docker image build by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/331
* Bump actions/setup-python from 6.2.0 to 6.3.0 in the actions group by @dependabot[bot] in https://github.com/huggingface/speech-to-speech/pull/332
* Update interruption handling docs to match current CancelScope logic by @A-Mahla in https://github.com/huggingface/speech-to-speech/pull/335
* Add realtime web demo app under demo/ by @A-Mahla in https://github.com/huggingface/speech-to-speech/pull/336
* Fix Windows installation error for faster-qwen3-tts by @mubashirsidiki in https://github.com/huggingface/speech-to-speech/pull/340
* Add star history chart to README by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/345
* Self-host star history chart by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/346
* Use GraphQL for star history data by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/347
* Make star history updates manual by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/348
* Add generated star history chart by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/349
* chore: update star history chart [skip ci] by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/350
* Retry API warmup before marking the node ready by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/353
* Use public Qwen3-TTS warmup API by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/355
* Prepare release 0.2.11 by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/362

## New Contributors
* @abhimanyus1997 made their first contribution in https://github.com/huggingface/speech-to-speech/pull/329
* @mubashirsidiki made their first contribution in https://github.com/huggingface/speech-to-speech/pull/340

**Full Changelog**: https://github.com/huggingface/speech-to-speech/compare/v0.2.10...v0.2.11