FAQ

Answers to common questions about gitmind: MCP memory, Claude and ChatGPT support, privacy, pricing, self-hosting and how it compares.

What is gitmind?

gitmind is an open-source memory for AI agents. It stores what Claude, ChatGPT, Claude Code and Codex learn as Markdown notes in a git repository you own, and serves the same notes to every app over MCP (Model Context Protocol). See the Introduction.

What is an MCP memory server?

MCP is the open protocol AI apps use to call outside tools. A memory server gives the agent tools to search, read and save notes, so it remembers things between conversations. gitmind provides these tools.

Which apps does gitmind work with?

claude.ai on the web, Claude Desktop and Claude mobile, ChatGPT, Claude Code and Codex, plus any other app that supports remote MCP servers with OAuth. See Connect your AI apps.

How do I give Claude Code persistent memory?

Add the endpoint with claude mcp add --transport http --scope user gitmind https://gitmind.app/mcp, or install the CLI and run gitmind setup for a synced local clone plus hooks that load your project’s notes at the start of each session. See CLI and local sync.

Can Claude and ChatGPT share the same memory?

Yes. Connect both to https://gitmind.app/mcp with the same account, and they read and write the same notes.

Is gitmind free?

Yes. gitmind.app is free with a GitHub sign-in, and the code is open source (AGPL-3.0), so you can also run your own.

Does gitmind work with my Claude Max or ChatGPT subscription?

Yes. gitmind is a connector, not a proxy: your app keeps talking to Claude or ChatGPT with your normal login and calls gitmind only to read and write notes.

Do I need GitHub?

You sign in with GitHub. Your notes live on gitmind.app (or your own deployment); a GitHub repository is only used if you choose CLI sync with gitmind setup instead of the cloud.

Can I edit notes myself?

Yes: on gitmind.app, in any editor on a CLI clone, or in Obsidian. Your edits sync like the agents’ do.

How is this different from built-in AI memory?

Built-in memory is per app and managed by its provider. gitmind is shared across every app and machine, stored as files you can read, edit, version and take with you, and shared with your team when you want. See How gitmind compares, including Basic Memory and Mem0.

Is my memory private?

Your personal workspace is readable only by you, and gitmind never sees your conversations: agents send it only what they choose to save or search. See Security and privacy.

Does it use semantic or vector search?

Yes, in gitmind cloud. Search ranks notes by meaning and by keywords together, so “how do we ship” finds a note about the deploy script, in any language. gitmind.app embeds each note’s passages with Cloudflare Workers AI (the multilingual Qwen3-Embedding model) when it changes, and keeps the vectors next to the keyword index: an index you can rebuild, never stored in your git repository. Filters by folder, tag, category and date apply to both.

The CLI searches its local clone by keyword, and adds the cloud’s semantic matches when your notes sync with gitmind cloud. With GitHub-only sync, search is keyword-only.

Does it work offline?

With the CLI, yes: agents search the local clone, and changes sync when you’re back online. The hosted endpoint needs a connection.

Can I leave and take my memory with me?

Yes. Every workspace is a git repository: clone it, or download it as a zip. The notes are plain Markdown that opens in Obsidian or any editor.

What happens if two agents write at the same time?

Both changes land. See How it works.