Show GN: Ravenclaw - An Open-Source System for Managing Work Context of AI Coding Agents
Key point
An open-source system that centrally manages and visualizes work context across AI coding agents.
Details
This is an open-source system designed to solve the context loss and difficulty of managing multiple projects that occur when using AI coding agents (Claude Code, Gemini CLI, etc.).
It leverages the MCP (Model Context Protocol) to centrally manage work context regardless of the type of agent. Even if an agent ends its session or is swapped out for another tool, the previous work state can be loaded as-is, allowing work to continue seamlessly.
The key features are as follows:
- Agent-centric workflow: Agents directly handle issue creation, status changes, and wiki writing, while humans monitor overall progress through a web UI and provide direction when needed.
- Project visualization: The web UI lets you grasp the Epic and issue structure and progress at a glance through a graph view.
- Context snapshots: Provides the ability to save work progress so it can be resumed immediately in the next session.
- Multi-agent support: Various agents such as Claude Code, Gemini CLI, and Codex can be used interchangeably within the same project environment.
It is built as a TypeScript-based monorepo, using Hono, PostgreSQL, React, and MCP Protocol as its tech stack. It is a self-hosted solution provided under the Apache 2.0 license.
This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.
Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.