Luthn: self-hosted shared memory for Codex and Claude Code

An open-source project called Luthn lets Codex and Claude Code share project memory by connecting through their hooks and MCP, collecting a bounded capsule of each agent's final response, then classifying and redacting it before making permitted context available to later tasks. Luthn runs on user-managed infrastructure with Docker and PostgreSQL and includes a local operator console for reviewing sensitive-access requests, aiming to let agents reuse project context while keeping private source material outside their default context.

I’d like to share Luthn , an open-source project for sharing project memory between AI agents. Luthn connects to Codex and Claude Code through their hooks and MCP. After a turn, it collects a bounded capsule of the agent’s final response, classifies and redacts it, and makes permitted context available for later tasks. When a new task starts, the agent can retrieve a small context pack. The project runs on infrastructure you manage, using Docker and PostgreSQL. It also includes a local operator console for reviewing sensitive-access requests. The goal is to let agents reuse project context while keeping private source material outside their default context. Repository: GitHub - JakobSung/Luthn: Self-hosted shared memory and safe context for AI agents. · GitHub https://github.com/JakobSung/Luthn I’m interested in feedback on both the design and the problem itself: