llms-install.md

Installing anki-mcp-server with LLMs

This guide provides step-by-step instructions for installing and configuring the anki-mcp-server MCP server using Large Language Models (LLMs).

Installation Steps

Preferred: Install via Desktop Extension (.mcpb)

1. Download the .mcpb bundle from the Releases page (or build locally with mcpb pack).
2. Open Claude Desktop → Settings → Extensions.
3. Drag the .mcpb file into the window and click Install.

To build the bundle locally:

npm install -g @anthropic-ai/mcpb
mcpb pack

This validates manifest.json and outputs an installable archive. See: Desktop Extensions: One-click MCP server installation for Claude Desktop.

Alternative: Manual configuration for Claude Desktop

Edit your Claude configuration file:

• MacOS: ~/Library/Application Support/Claude/claude_desktop_config.json
• Windows: %APPDATA%/Claude/claude_desktop_config.json

{
  "mcpServers": {
    "anki": {
      "command": "npx",
      "args": ["--yes", "anki-mcp-server"]
    }
  }
}

Configuration for Cline

Add the anki server to your cline_mcp_settings.json

{
  "mcpServers": {
    "anki": {
      "command": "npx",
      "args": ["--yes", "anki-mcp-server"]
    }
  }
}

Verification

To verify the installation:

1. Ensure Anki is running
2. Start Claude
3. Try creating a new note with the following prompt:

Create an Anki card in the "Default" deck with:
Front: What is the capital of France?
Back: Paris

Troubleshooting

1. "Cannot connect to AnkiConnect" error:

   • Ensure Anki is running
   • Check if AnkiConnect is properly installed
   • Restart Anki and try again

2. "Permission denied" when running the server:

   • Ensure the build file is executable:

   chmod +x build/index.js

3. "Deck not found" error:

   • Create the deck manually in Anki first
   • Check for exact deck name spelling

Testing

To run the test suite:

npm test

This will verify:

• Server initialization
• AnkiConnect communication
• Note creation/deletion
• Deck operations

For development with auto-rebuild:

npm run watch