@jarvis563

stable-browser replaces openclaw's flaky extension relay with direct chrome devtools protocol connections, eliminating websocket 403 errors and disconnects via one-time setup script and dedicated debug profile.

percept-voice-cmd detects wake words in ambient speech and routes commands (email, text, reminders, search, calendar, notes) to openclaw agents. uses two-tier intent parsing with speaker authorization via dashboard or json config.

percept-listen captures ambient audio from wearable devices (omi pendant or apple watch), transcribes locally, and streams speaker-tagged conversation data to openClaw agents via webhook and local sqlite storage.

browser-audio-capture enables agents to ingest audio from any chrome tab via a local extension or cli tool, streaming pcm16 frames to a configurable http endpoint for transcription or analysis.

Continuously captures and summarizes ambient conversations to build a local knowledge graph for context-aware assistance without explicit commands.

Give your OpenClaw agent context from meetings captured by Zoom, Granola, or Omi wearables. Search transcripts, extract action items, identify speakers, and...