livekit-agents@1.8.0 livekit-agents@1.8.1
fosrl/pangolinSep 5, 2026 → Sep 10, 20266 days between releases
What's New
Introduces DuplexModel class for simultaneous speech and listening with OpenAI GPT-Live
New Features
- New DuplexModel class for simultaneous speech and listening
- OpenAI GPT-Live model as the first duplex model implementation
- New Muse Voice streaming STT plugin
- Phonic watermarking configuration option
Improvements
- Telemetry improvements to describe option objects in session reports
- Isolated ruff dependencies in CI
- TTS pool invalidation fixes
Bug Fixes
- OpenAI TTS prewarm fix to prevent replacing in-flight requests
- Drained unused audio forwarding buffer
- Removed commented-out Deepgram V2 STT keepalive stub
Upgrade Notes
No breaking configuration changes are required. The new DuplexModel class is an addition for full-duplex scenarios, while existing telemetry remains compatible.
The Two Releases
livekit-agents@1.8.0
Sep 5, 2026· 2 months ago
This release focuses on OpenTelemetry changes, migrating spans, metrics, and logs to GenAI semantic conventions and moving PII filtering in-process. It includes significant breaking changes to telemetry data structures.
Full release noteslivekit-agents@1.8.1
Sep 10, 2026· 1 month ago
A major feature release introducing full-duplex speech model support via the new `DuplexModel` class, starting with OpenAI's GPT-Live model. The release also includes various bug fixes and new telemetry capabilities.
Full release notes