0.11.68

Automattic/harper0.11.68Jul 15, 2026by halajohn

AI Summary

This release focuses on enhancing speech recognition capabilities with improved Soniox ASR handling and introduces support for new Text-to-Speech providers like Mistral and Gradium. It also includes updates for avatar rendering, connection state management, and dependency optimizations for Python extensions.

Key Highlights

  • Enhanced Soniox ASR with improved sentence termination handling
  • Added support for Mistral Voxtral and Gradium TTS extensions
  • Improved avatar handling with final data for EOF and enabled Opus by default
  • Updated Aliyun ASR to support dashscope 1.26.0+
  • Optimized dependencies by scoping numpy and allowing newer websockets versions

New Features

  • Soniox ASR sentence termination handling
  • Mistral Voxtral TTS extension
  • Gradium TTS support
  • Avatar EOF finalize data
  • Spatius Opus audio codec
  • Aliyun ASR dashscope support
  • Websocket version updates

Full Release Notes

## What's Changed
* feat: enhance Soniox ASR with sentence termination handling by @diyuyi-agora in https://github.com/TEN-framework/ten-framework/pull/2196
* chore: add pyproject.toml files for python extensions by @anexpn in https://github.com/TEN-framework/ten-framework/pull/2197
* feat: add Mistral Voxtral TTS extension by @TiagoAgora in https://github.com/TEN-framework/ten-framework/pull/2194
* feat(tts): add gradium tts support by @BenWeekes in https://github.com/TEN-framework/ten-framework/pull/2193
* feat(avatar): use finalize data for EOF by @YiminW in https://github.com/TEN-framework/ten-framework/pull/2199
* feat(spatius): enable opus by default for spatius avatar by @3DRX in https://github.com/TEN-framework/ten-framework/pull/2201
* feat(aliyun_asr): support dashscope 1.26.0+ by @anexpn in https://github.com/TEN-framework/ten-framework/pull/2204
* feat: allow newer websockets versions across Python extensions by @anexpn in https://github.com/TEN-framework/ten-framework/pull/2205
* feat: scope numpy deps to extensions that need them by @anexpn in https://github.com/TEN-framework/ten-framework/pull/2207
* feat: openai tts and rime tts update metrics error connection status by @YiminW in https://github.com/TEN-framework/ten-framework/pull/2211
* feat(asr): add connection status validation tests and enhance vendor … by @diyuyi-agora in https://github.com/TEN-framework/ten-framework/pull/2209
* feat: rime_tts update connection state report by @YiminW in https://github.com/TEN-framework/ten-framework/pull/2212
* update trendshift badge and banner by @bluemotional in https://github.com/TEN-framework/ten-framework/pull/2213
* fix: mistarl tts pop sample_rate in reqeust by @YiminW in https://github.com/TEN-framework/ten-framework/pull/2216
* fix: enhance error handling and callback mechanisms by @diyuyi-agora in https://github.com/TEN-framework/ten-framework/pull/2218
* fix: mistral tts audio frame by @YiminW in https://github.com/TEN-framework/ten-framework/pull/2219
* chore: bump version to 0.11.68 by @halajohn in https://github.com/TEN-framework/ten-framework/pull/2220

## New Contributors
* @TiagoAgora made their first contribution in https://github.com/TEN-framework/ten-framework/pull/2194
* @bluemotional made their first contribution in https://github.com/TEN-framework/ten-framework/pull/2213

**Full Changelog**: https://github.com/TEN-framework/ten-framework/compare/0.11.67...0.11.68