Xepha — an AI memory layer and CLI for coding agents

I’m building Xepha: a memory + coordination layer for AI coding agents.

The idea is to help tools like Claude Code / Codex-style workflows keep better project context through a simple repo-based structure, persistent memory, and cleaner orchestration for developers.

Repo: https://github.com/francozeta/xepha

I’d love feedback on:

  • the concept
  • the CLI / DX
  • the architecture
  • how to make it more useful for real-world agent workflows

Any thoughts, ideas, or suggestions are very welcome.