Self-hosted MCP server that lets Claude, ChatGPT, Meta Muse and any MCP client read and write one Markdown vault (Obsidian-compatible). Go, OAuth 2.1 with passkeys, no shell, no hard deletes.
A self-hosted MCP server that lets Claude, ChatGPT, Meta Muse, and any other MCP client read and write the same folder of Markdown notes. The notes are the memory; the assistant behind them is replaceable.
Every assistant keeps its own memory, and that memory is locked to the vendor. Move from one app to another, or use two at once (one on the desktop, another on the phone), and each one knows a different slice of you. When the app changes, the memory goes with it.
cortex-mcp turns that around. Your memory is a plain folder of .md files that you own, that opens in Obsidian or any editor, and that you can sync, back up, and read without any AI at all. The server puts that folder on the internet behind a proper login, and every assistant that speaks MCP connects to it. Swap the assistant whenever you like; the notes stay.
- Exposes one vault to every MCP client over Streamable HTTP, at a single URL such as https://notes.example.com/mcp. - Gives assistants twelve tools to read, search, and edit notes (table below). There is no shell, no git, and no hard delete: a deleted note goes to .trash/. - Protects edits with a per-note version, so a stale edit fails with notechanged instead of overwriting a newer one, and records every write in a log with the name of the app that made it. - Lets assistant apps connect with OAuth 2.1 (claude.ai, ChatGPT, Meta Muse, Claude Code). Each connection is approved by you on a login page, with a passkey, an authenticator code, or a recovery code. - Issues Bearer tokens for CLI agents and scripts that do not do OAuth. - Sends your own instructio
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/JoseJimenez-M--cortex-mcp.html)
Turn scattered notes, docs and transcripts into a queryable Markdown wiki — an LLM knowledge-base compiler with MCP access, no embeddings, self-hosted.
rwnalds/engramSelf-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
do-md/homekbLocal-first personal Markdown knowledge base with semantic search, AI Q&A, MCP, desktop/web clients, and a self-hostable zero-data-at-rest relay.
menot-you/n-memoryMemory your agent can trust — every answer carries its source; when it doesn't know, it says so. Hermetic, local, MCP.
systempromptio/systemprompt-demoAI Governance Infrastructure — open gateway demo. Self-hosted systemprompt.io gateway in a single Rust binary: policy, audit trails, and cost attribution on every AI request, with the Systemprompt Bri
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
The top new MCP servers of the week, every Monday. No spam, unsubscribe anytime.