data-ai MCP Server
Local-first markdown notes with a first-class MCP server. 16 tools + 2 resources for Claude Code, Claude Desktop, Cursor and any MCP agent. Per-folder permissions, semantic RAG via LanceDB, single Bun binary. AGPL.
Discovered via github-topic:mcp-server and last synced 3mo ago.
1. Install the package
npm install brain.md
—
`read`
`2048`
`read`
`read`
`nvidia/llama-nemotron-embed-vl-1b-v2:free`
Proprietary
`read`
`brain.md` (or anything you like)
Value
Logseq
`read`
**AGPL-3.0**
`read`
`read`
+----------------+
[`brain-md-linux-arm64`](https://github.com/mi4uu/brain.md/releases/latest/download/brain-md-linux-arm64)
[`brain-md-windows-x64.exe`](https://github.com/mi4uu/brain.md/releases/latest/download/brain-md-windows-x64.exe)
Path
`/api/rag/status`
`read`
none
`3000`
—
OpenAI-compatible
Folder perm
none
Env var
—
✓
`/api/rag/test`
`read`
`write`
Default
`1`
✓
none
`XDG_DATA_HOME`
`write`
`15000`
e.g. `nomic-embed-text`
none
`XDG_CONFIG_HOME`
Architecture
`text-embedding-3-small`
`read`
XDG