Gryph is a local-first audit trail for AI coding agents. It hooks into your agents, logs every action to a local SQLite database, and gives you powerful querying capabilities to understand, review, and debug agent activity.
Why Gryph?
- Transparency - See exactly what files were read, written, and what commands were run
- Audit Trail - See exactly what your AI coding agent did
- Debugging - Replay sessions to understand what went wrong
- Privacy - All data stays local. No cloud, no telemetry
Copilot says:
AI-generated
Finally, a flight recorder for your AI coding agents — every file read, every command run, every edit made gets logged locally so you can actually see what Claude or Cursor did while you werent looking.
Key features:
- 📜 Logs every agent action to a local SQLite database for full transparency
- 🔍 Query and filter sessions to debug what went wrong or review changes
- 🔌 One-command install hooks for Claude Code, Cursor, Gemini CLI, and more
- 🔒 Privacy-first with all data staying local and zero telemetry
This summary was generated by GitHub Copilot based on the project README.