livekit-agents@1.6.2

What's Changed

• feat(assemblyai): add universal-3-5-pro (now default) and Voice Focus streaming params by @dlange-aai in #6119
• feat: add Gemini 3.1 flash TTS support, implement streaming response … by @thorwebdev in #6134
• (google): fix formatting by @tinalenguyen in #6142
• chore(examples): update livekit-agents version to >=1.6 in requirements.txt by @Bobronium in #6150
• fix(openai): default reasoning_effort to 'none' for mini/nano models by @sarvesh4396 in #6148
• feat(assemblyai): add streaming mode (latency/accuracy preset) param by @dlange-aai in #6156
• feat(fishaudio): add speed/volume prosody params and default to s2.1-pro by @cshape in #6159
• docs: add missing google_gemini_tts.py example to voice agents README by @detail-app[bot] in #6155
• feat(soniox): support stt-rt-v5 with endpoint_sensitivity option by @mihafabcic-soniox in #6126
• fix(xai): expose TTS websocket options by @nightcityblade in #6095
• (phonic) hotfix: add a config that lists tools with forbid-speech-after-tool-call by @arunwpm-work in #6143
• fix(workflows): normalize two-digit years in GetDOBTask by @he-yufeng in #6124
• fix(azure): emit STT recognition usage metrics by @nightcityblade in #6161
• (phonic): reuse ws connection across handoffs by @tinalenguyen in #6163
• (google): surface context exhaustion error by @tinalenguyen in #6144
• (inference): add assembly ai model by @tinalenguyen in #6162
• docs(phonic): add forbid_speech_after_tool_call to README config table by @detail-app[bot] in #6168
• feat(phonic): handle user text input by @tinalenguyen in #6169
• livekit-agents@1.6.2 by @github-actions[bot] in #6170

New Contributors

• @thorwebdev made their first contribution in #6134
• @sarvesh4396 made their first contribution in #6148

Full Changelog: https://github.com/livekit/agents/compare/livekit-agents@1.6.1...livekit-agents@1.6.2