affine
MCP server from DAWNCR0W/affine-mcp-server
claude mcp add --transport stdio dawncr0w-affine-mcp-server affine-mcp
How to use
This MCP server integrates AFFiNE workspaces and documents with AI assistants. It exposes AFFiNE resources over MCP using the standard AFFiNE MCP server tooling, primarily via a local stdio channel by default, with an optional HTTP /mcp endpoint for remote deployments. The server supports 43 tools for workspace and document management, including creating and updating workspaces, listing and retrieving documents, publishing and revoking documents, markdown import/export, tag handling, database block workflows, comment management, version history, user and token management, notifications, and blob storage operations. To run locally, install the global affine-mcp-server CLI and start the server; the server will read credentials from the AFFiNE MCP config stored by the affine-mcp CLI. For remote deployments, configure the HTTP transport and provide the necessary AFFINE_BASE_URL and authentication tokens or cookies so clients can connect over HTTP at the /mcp path.
How to install
Prerequisites:
- Node.js 18+ and npm
- Access to an AFFiNE instance (self-hosted or cloud) with valid credentials or API token
Installation steps:
- Install the CLI globally (recommended):
npm i -g affine-mcp-server
- Start the server (stdin mode):
affine-mcp
This runs the MCP server over stdio. Alternatively you can run ad-hoc via npx without installation:
npx -y -p affine-mcp-server affine-mcp -- --version
- Optional: verify installation and credentials by performing a login to populate ~/.config/affine-mcp/config. The CLI login stores credentials securely for the MCP server to read automatically.
Configuration via CLI (interactive login):
affine-mcp login
Follow prompts to connect to your AFFiNE instance and generate an API token for MCP access. Credentials are stored in a secure config file and read automatically by the MCP server.
Configuration via environment variables (alternative):
- AFFINE_BASE_URL is required
- Auth: AFFINE_API_TOKEN or AFFINE_COOKIE or AFFINE_EMAIL + AFFINE_PASSWORD
- Optional: AFFINE_GRAPHQL_PATH, AFFINE_WORKSPACE_ID, AFFINE_LOGIN_AT_START
Example using environment variables (replace with real values):
export AFFINE_BASE_URL=https://app.affine.pro
export AFFINE_API_TOKEN=ut_xxx
export AFFINE_GRAPHQL_PATH=/graphql
# Start the server in stdio mode or configure your MCP client to connect to the CLI's stdio channel
Additional notes
Tips:
- Prefer affine-mcp login or AFFINE_API_TOKEN for zero-latency startup.
- If using a password in shells, consider quoting to avoid shell history expansion issues.
- For remote HTTP deployments, set MCP_TRANSPORT=http and protect /mcp with a strong token (AFFINE_MCP_HTTP_TOKEN) and origin controls (AFFINE_MCP_HTTP_ALLOWED_ORIGINS).
- When using Cloud AFFiNE, avoid Cloudflare blocks by using token-based authentication or the login flow which handles token creation automatically.
- The README notes 43 tools are available; familiarize yourself with the MCP SDK documentation to leverage WebSocket-based document editing and workspace management features.
- The CLI wrapper ensures Node runs the ESM entrypoint to prevent shell misinterpretation of JS modules.
Related MCP Servers
iterm
A Model Context Protocol server that executes commands in the current iTerm session - useful for REPL and CLI assistance
mcp
Octopus Deploy Official MCP Server
furi
CLI & API for MCP management
editor
MCP Server for Phaser Editor
DoorDash
MCP server from JordanDalton/DoorDash-MCP-Server
mcp
MCP сервер для автоматического создания и развертывания приложений в Timeweb Cloud