Connect your models and chats together with one central git memory. Plain Markdown facts in a git repo you own — Claude Code, ChatGPT, Cursor, Codex and anything that uses APIs, Integrations or MCPs r
View on GitHub ↗Connect your models and chats together with one central git memory.
Your AI assistants forget everything between sessions. This gives them one memory, in a git repo you own, that every one of them can read.
Write a fact in Claude Code. Ask ChatGPT about it. Get it back.
That makes a repo. Push it to a private GitHub repo, connect your tools, and they share one memory. No server, no database, no account with anyone, no subscription. The memory is markdown files in your own git.
Because nearly every AI client can already read one, and most can write to one.
The table below is read from vendor documentation on 21 September 2026. Claude Code is the one row tested on a real repo. Vendors add and drop connector support often, so check your own client before you build on a row.
Client Read Write --------- Claude Code, Codex, Cursor, Gemini CLI, Copilot, Windsurf, Aider Yes Yes ChatGPT Yes Yes Claude, Gemini, Grok, Mistral Vibe, Perplexity, M365 Copilot Yes, via connector Yes, via connector Amazon Q, DeepSeek Harness, Qwen Chat Yes, via connector Yes, via connector
"Via connector" means you point the client at your repo with whatever it supports: a GitHub integration, the GitHub API, or GitHub's hosted MCP server. All three read the same folder. You host nothing.
Two with no connector surface in their documentation on that date: Meta AI and the DeepSeek consumer app.
memory/ holds what is true now. memory/archive/ holds what used to be true.
From the project README.
Add the radar badge to your README — it shows your project was picked up by MCP Radar and links to this page:
[](https://mcp.liqiwa.com/s/JMacCyber--jmac-memory-os.html)
Self-hosted MCP server + dashboard giving Claude Code, Cursor & Hermes agents shared long-term memory over a git-backed markdown vault. Agent-native, Obsidian-compatible second brain — markdown RAG, n
gonzaroman/IHMT-MEMORYLong-term memory for AI coding agents (Claude Code, Codex, opencode) as a tree of plain files
vinodborole/okf-kitTurn any website into a portable, agent-ready Open Knowledge Format (OKF) bundle — no LLM required. Crawl, sync, chat offline, and serve docs to coding agents over MCP.
okf-memory/okf-agent-memoryGit-native persistent memory for AI coding agents. Implements Google OKF v0.2 with sub-300µs in-memory BM25 search, embedded MCP server, and progressive disclosure. Slashes token bloat by 80% with zer
Kerneta/daidocsOpen plain-text file format for AI memory. Your assistant's long-term memory as .dai files on your disk: readable by Claude, GPT, Gemini, Cursor, local models and grep (all LLM models work). MCP serve
Consensus-NLP/consensus-mcpOfficial MCP server for Consensus. Search 220M+ peer-reviewed research papers from Claude, ChatGPT, Cursor, and any MCP client.
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.