AI Agent Local Q&A & Transcript (v1)
Conversational Q&A grounded in your notes, with brain source indicator, Hold-to-Talk, New Chat, and Transcript viewer.
Grounded in your notes
AI Agent Voice Assistant
Hold-to-talk Q&A grounded in your Obsidian vault, worklogs, and MemoryHive summaries — running on your localhost by default.
- Runs against your local model by default; explicit per-session approval before any cloud model
- Retrieves relevant snippets from your personal notes and recent meeting transcripts
- On-device Transcript viewer lets you read full responses or replay spoken TTS
- Automatic road gateway failover keeps your AI Agent reachable away from home
Default BrainLocalhost LLM
TTSPiper Voice Packs
PrivacyPer-Chat Cloud Consent
Screen-Only Captures (368×448)
What this screen shows