anki

This MCP server provides a bridge between a locally running Anki desktop app (with the Anki-Connect add-on installed) and your automation or tooling. It exposes tools to query cards from Anki (via Anki-Connect) and to perform actions such as creating new cards or updating existing ones. The available tools are: get_due_cards, get_new_cards, add_card, and update_cards. You can use these to build workflows that review, study, or populate your Anki deck without manually interacting with the UI. When integrated through Claude Desktop or other MCP clients, the server will read input from stdin and return results over stdout, enabling scripted interactions with your Anki collection.

Prerequisites:

• Node.js and npm installed on your system
• Anki desktop application installed with the Anki-Connect add-on configured and running

Install and build:

1. Clone or download the MCP server repository
2. Install dependencies

   npm install
   

3. Build the server

   npm run build
   

4. (Optional) Run in development mode with auto-rebuild

   npm run watch
   

Usage note: Ensure Anki is running and Anki-Connect is enabled so the MCP server can communicate with Anki to fetch or modify cards.

Tips and caveats:

• MCP servers communicate over stdio; for debugging, consider using the MCP Inspector tool to view and troubleshoot protocol interactions.
• The server exposes functions to get due/new cards and to add or update cards in your default deck. Use get_due_cards and get_new_cards to paginate results by specifying the num parameter.
• In the Claude Desktop config example, the server is referenced via a path to the built entry (build/index.js). Update this path to reflect your actual deployment location.
• Ensure Anki-Connect is reachable (the add-on provides a local HTTP API); if you run into connection issues, verify that Anki is running and the add-on is enabled, and that no firewall blocks localhost connections.
• If you customize environments or paths, consider adding environment variables as needed and documenting them in env within your deployment config.