OpenCode Voice 0.0.3

Added\n- Handy local transcription provider using installed Handy models.\n- Configurable Handy executable resolution for AppImages and GUI-launched OpenCode.\n\n## Changed\n- Hardened release artifact verification and production dependency auditing.

OpenCode Voice 0.0.2

OpenCode Voice 0.0.2 expands installation, microphone selection, credential reuse, and prompt status behavior.

Added

• Guided local Whisper model installation with progress, cancellation, retry, minimization, and recovery.
• Linux microphone discovery and explicit device selection across PulseAudio, PipeWire, and ALSA capture backends.
• Configurable microphone level display.
• Reuse of OpenCode-managed OpenAI API keys when available.

Changed

• Extracted voice sessions, audio capture, and transcription providers into focused runtimes with expanded regression coverage.
• Limited the validated release environment to Linux x64 with Node 22.22.2 and OpenCode 1.18.3.
• Documented real microphone capture as experimental while hardware compatibility testing remains incomplete.
• Refined voice status rendering across home and session prompts.

Fixed

• Made runtime activation transactional and serialized managed asset mutations.
• Improved recording cleanup, installer cancellation, and failure recovery.
• Preserved native prompt alignment and prevented updates to destroyed status renderables.

Support

This prerelease supports Linux x64 with Node 22.22.2. Other operating systems, architectures, and JavaScript runtimes are not release-supported.

OpenCode Voice 0.0.1

Initial Linux-only prerelease of OpenCode Voice.\n\nSupports Linux x64 and arm64 with local Whisper, OpenAI, and ElevenLabs transcription. Release artifacts are produced and verified by the pinned CI workflow.