Starstruk Logo MarkStarstruk

v0.2.10

Mesh-LLM/mesh-llmv0.2.10Jun 11, 2026by andimarafioti

AI Summary

Focuses on infrastructure improvements including concurrent session pools, speculative turn tracking, and fixes for Paraformer transcription and speculative VAD handling.

Key Highlights

  • Added `--num_pipelines` pool for managing concurrent realtime sessions.
  • Implemented speculative turn revisions to improve tracking logic.
  • Fixed Paraformer progressive transcription events.
  • Improved speculative VAD reopen and continuation handling.

New Features

  • Concurrent session pool management
  • Speculative turn revisions
  • Paraformer transcription fixes
  • Speculative VAD improvements

Full Release Notes

## What's Changed
* Add --num_pipelines pool for concurrent realtime sessions by @A-Mahla in https://github.com/huggingface/speech-to-speech/pull/282
* Add speculative turn revisions by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/255
* Bump the actions group with 5 updates by @dependabot[bot] in https://github.com/huggingface/speech-to-speech/pull/293
* Normalize Qwen3-TTS language aliases by @kamjin3086 in https://github.com/huggingface/speech-to-speech/pull/300
* [codex] Fix TTS benchmark input message by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/303
* Bump the actions group with 2 updates by @dependabot[bot] in https://github.com/huggingface/speech-to-speech/pull/301
* Fix Paraformer progressive transcription events by @kamjin3086 in https://github.com/huggingface/speech-to-speech/pull/299
* Improve speculative VAD reopen and continuation handling by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/307
* Fix tool-only realtime response completion by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/306
* Shorten voice prompt and improve tool lead-ins by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/305
* Prepare 0.2.10 release by @andimarafioti in https://github.com/huggingface/speech-to-speech/pull/311

## New Contributors
* @dependabot[bot] made their first contribution in https://github.com/huggingface/speech-to-speech/pull/293
* @kamjin3086 made their first contribution in https://github.com/huggingface/speech-to-speech/pull/300

**Full Changelog**: https://github.com/huggingface/speech-to-speech/compare/v0.2.9...v0.2.10