v0.8.0 - Speech-to-Text & SSE Improvements
New Features
- Speech-to-Text (STT): Full voice input implementation with recording overlay UI
- Support for browser-native Web Speech API
- Support for OpenAI Whisper backend
- New settings panel for configuring STT provider, language, and Whisper-specific options
- Recording aborts automatically on input clear
Bug Fixes
- Git log dates: Fixed invalid date display in commits tab by using Unix timestamps
- Session timestamps: Fixed last active timestamp showing incorrect values
- Session display: Fixed NaN showing on mobile session cards
- SSE stability:
- Fixed client cleanup on disconnect
- Fixed duplicate Transfer-Encoding headers
- Added proxy-compatible headers for better tunnel support
- Removed debug test endpoint
Improvements
- Textarea auto-height: Use CSS
field-sizingproperty for cleaner auto-resize behavior
Housekeeping
- Removed CHANGELOG.md - all release notes are now maintained exclusively in GitHub Releases
Upgrading
If you're using Docker, pull the latest image:
docker pull ghcr.io/chriswritescode-dev/opencode-manager:latestFull changelog: v0.7.8...v0.8.0