FastMCP FastMCP
Sign In
A

Anki MCP Server

An MCP server that enables AI assistants to interact with the Anki flashcard application for studyin...

35 views
0 installs
Updated Feb 10, 2026
Not audited
API Development Developer Tools
Tools I Recommend
ClaudeKit ClaudeKit Sponsor
An MCP server that enables AI assistants to interact with the Anki flashcard application for studying, deck management, and note creation. It supports natural language interaction for reviewing cards, searching content, and managing media files across local and remote environments.

  1. Click the "Install Now" button and open the FastMCP connection interface.

    • You will enter the ENV names/values in that FastMCP connection form — fill each field there (not in a local shell).

  2. Required/optional ENV fields to add in the FastMCP connection interface (what to put and how to obtain each value):

    1. ANKI_CONNECT_URL — (what to put)

      • Default: http://localhost:8765
      • How to confirm or change it:
        1. Open Anki → Tools → Add‑ons → select AnkiConnect → click Config.
        2. In the JSON config, read webBindAddress and webBindPort and build the URL as http://:. If webBindAddress is 0.0.0.0 (or you exposed it), use the host IP or keep localhost for local clients. (github.com)

    2. ANKI_CONNECT_API_KEY — (secret token; optional)

      • What it is: an API key you can enable in AnkiConnect so external clients authenticate.
      • How to create/set it (in Anki):
        1. Open Anki → Tools → Add‑ons → select AnkiConnect → click Config.
        2. In the JSON config, set the apiKey value to a strong secret string, e.g. "apiKey": "my-secret-key" (replace with your chosen secret).
        3. Save and restart Anki.
        4. Copy that exact secret string into the FastMCP ANKI_CONNECT_API_KEY field.
      • If you leave apiKey null in AnkiConnect, leave the FastMCP ANKI_CONNECT_API_KEY field empty. (github.com)

    3. ANKI_CONNECT_API_VERSION — (optional)

      • Default: 6. If you have no special requirement, put 6 (or leave the FastMCP field blank if the interface supplies a default).

    4. ANKI_CONNECT_TIMEOUT — (optional)

      • Default: 5000 (milliseconds). Set a numeric timeout in the FastMCP field if you need a longer/shorter request timeout.

    5. READ_ONLY — (optional)

      • If you want the MCP server to block any write operations, set this ENV to true (or 1). Otherwise set false or leave unset.

    6. HTTP-mode ENV fields (only if you plan to run HTTP mode on the remote server instead of STDIO):

      • PORT — HTTP port (default 3000).
      • HOST — Bind address (default 127.0.0.1).
      • ALLOWED_ORIGINS — CORS allowed origins (comma-separated).
      • LOG_LEVEL — e.g., info, debug.
      • Only populate these in FastMCP if you will run the server in HTTP mode; otherwise STDIO clients can ignore them.

    7. (Optional) ngrok auth token — only needed if you plan to use the integrated --ngrok tunnel

      • How to get it: sign in to your ngrok dashboard and copy your auth token.
      • How to apply it (one-time, on your machine): run in a terminal: ngrok config add-authtoken <YOUR_TOKEN> (or follow the ngrok dashboard instructions). After that you can start the server with --ngrok and it will create a public tunnel. (ngrok.com)

  3. Example quick checklist (do these in order):

    1. In Anki: Tools → Add‑ons → AnkiConnect → Config → confirm/set webBindAddress/webBindPort and (optionally) apiKey → Save → Restart Anki. (github.com)
    2. Open FastMCP connection interface (Install Now).
    3. Paste ANKI_CONNECT_URL (e.g., http://localhost:8765) into the FastMCP ANKI_CONNECT_URL field.
    4. Paste your AnkiConnect API key (if you set one) into the FastMCP ANKI_CONNECT_API_KEY field.
    5. Fill other optional fields (API_VERSION, TIMEOUT, READ_ONLY, PORT/HOST if using HTTP).
    6. Save the FastMCP connection and start the MCP server from your MCP client; verify by visiting your ANKI_CONNECT_URL (e.g., http://localhost:8765) or attempting a simple tool call.

  4. How to verify the connection after saving:

    • If AnkiConnect is running locally, open the ANKI_CONNECT_URL in a browser (http://localhost:8765) or run a small POST to the URL to call the "version" action; the service should respond and your MCP client should be able to list decks. (If you set an API key, your client must supply the same key.) (quizcraft.org)

  5. Troubleshooting tips

    • If FastMCP reports connection refused: confirm Anki is running and AnkiConnect is enabled; confirm webBindAddress and webBindPort in the AnkiConnect config; ensure firewall/host networking allows the connection. (github.com)

Done — fill the values into the FastMCP connection interface form (opened via Install Now) following the steps above.

Quick Start

View on GitHub

More for API Development

View All →

More for Developer Tools

View All →

Similar MCP Servers

Report Issue

Thank you! Your issue report has been submitted successfully.