Audio Capsule Recorder v0.6.2

Audio Capsule Recorder v0.6.2

Highlights

• Native-feeling recorder surface for Obsidian notes
• Progressive manual actions after each recording
• Split desktop/mobile runtime profiles for synced vaults
• Local or cloud AI post-process support
• Kimi upgraded to kimi-k2.6
• Kimi-safe compatibility mapping to avoid invalid temperature=0 requests

What this release contains

Core recording workflow

• In-note recorder UX for empty notes
• Compact floating recorder entry for non-empty notes
• Native Obsidian audio file insertion
• Progressive per-recording action blocks:
  • transcript + AI post-process
  • transcript only
  • post-process later

Reliability features

• single-flight lock per recording
• repair command for broken historical action markers
• explicit pending / failure / retry states
• split desktop/mobile configuration for synced vaults

Runtime support

Speech-to-text:

• Groq Whisper Large V3
• Groq Whisper Turbo
• SiliconFlow SenseVoice
• OpenAI-compatible STT
• localhost / custom STT

AI post-process:

• local or LAN OpenAI-compatible HTTP text services
• Kimi kimi-k2.6
• Volcengine Coding Plan
• Zhipu GLM Coding
• MiniMax

Mobile notes

• explicit mobile recording screen policy
• best-effort wake lock mode
• system-default power-saving mode
• manual lock-screen continuation still depends on current Obsidian mobile platform limits

Verification baseline

• npm run build
• npx tsc --noEmit

Welcome

Welcome to use the project, open issues, and build your own secondary development on top of it.