Anki MCP Server

Transform Claude into your AI-powered Anki assistant

Features • Installation • Usage • Contributing

---

What is Anki MCP?

A complete Model Context Protocol (MCP) server that connects Claude AI with Anki, enabling:

• AI-powered flashcard generation from any text, PDF, or webpage
• Smart review analytics and learning insights
• Automatic card improvements using spaced repetition science
• Natural language control of your entire Anki collection

MCP allows Claude to interact directly with Anki, making spaced repetition learning effortless.

Features

55 MCP Tools across 10 categories

Category	Tools	What you can do
Generation (Priority)	6 tools	Generate cards from text, create cloze deletions, improve existing cards
Notes	8 tools	Create, update, search, and manage notes
Decks	5 tools	Organize your collection with deck management
Cards	6 tools	Find, suspend, move cards with precision
Statistics	7 tools	Analyze retention, predict workload, get insights
Review	6 tools	Smart review sessions with AI suggestions
Models	4 tools	Manage note types and templates
Media	4 tools	Handle images, audio, and video
Sync	3 tools	Sync with AnkiWeb seamlessly
Import/Export	6 tools	Markdown, CSV, JSON support

Additional Features

• 9 MCP Resources: Real-time access to decks, stats, and cards
• 10 Smart Prompts: Guided workflows for card creation and review
• Pydantic Models: Type-safe data validation
• Async HTTP: Lightning-fast AnkiConnect integration
• Unit Tests: Reliable and tested codebase

Quick Start

Prerequisites

1. Anki running with AnkiConnect add-on (code: 2055492159)
2. Python 3.11+
3. Claude Desktop

Install AnkiConnect

1. Open Anki → Tools > Add-ons > Get Add-ons
2. Enter code: 2055492159
3. Restart Anki

Installation

# Clone the repository
git clone https://github.com/guepardlover77/anki-mcp.git
cd anki-mcp

# Create virtual environment
python -m venv venv
source venv/bin/activate  # Linux/Mac
# or
venv\Scripts\activate  # Windows

# Install dependencies
pip install -e .

# Or with extras for PDF and web
pip install -e ".[all]"

Claude Desktop Configuration

Add to your Claude Desktop config file (claude_desktop_config.json):

{
  "mcpServers": {
    "anki": {
      "command": "python",
      "args": ["-m", "anki_mcp"]
    }
  }
}

Environment Variables

Variable	Description	Default
ANKI_MCP_HOST	AnkiConnect host	localhost
ANKI_MCP_PORT	AnkiConnect port	8765
ANKI_MCP_API_KEY	AnkiConnect API key (optional)	-
ANKI_MCP_TIMEOUT	HTTP timeout in seconds	30
ANKI_MCP_DEBUG	Debug mode	false

MCP Resources

• anki://decks - List all decks
• anki://decks/{name} - Deck details
• anki://decks/{name}/due - Due cards for a deck
• anki://models - List note types
• anki://models/{name} - Model details
• anki://tags - All tags
• anki://stats/today - Today's statistics
• anki://notes/{id} - Note details
• anki://cards/{id} - Card details

Usage Examples

AI-Powered Card Generation

You: "Generate flashcards from this article about Python decorators"

Claude will:
- Extract key concepts
- Create Q&A pairs
- Add to your chosen deck
- Apply best practices automatically

Smart Analytics

You: "Analyze my weak areas in the Spanish deck"

Claude provides:
- Retention analysis
- Difficult card patterns
- Personalized study recommendations
- Predicted review workload

Batch Operations

You: "Import these 50 vocabulary words from CSV"

Claude handles:
- Format detection
- Duplicate checking
- Tag organization
- Progress reporting

Natural Language Commands

• "Create a deck for learning Japanese"
• "Find all suspended cards in my collection"
• "Export my French deck to Markdown"
• "Show me today's review statistics"
• "Generate cloze cards from this text about photosynthesis"

Development

# Install development dependencies
pip install -e ".[dev]"

# Run tests
pytest

# Check linting
ruff check src/

# Type checking
mypy src/

Verify Connection

python -m anki_mcp --check

Contributing

We welcome contributions! Check out our Contributing Guide to get started.

Ways to Contribute

• Report bugs or request features via Issues
• Submit PRs for new features or bug fixes
• Improve documentation with examples and tutorials
• Star the repo to show support
• Share with the Anki and MCP communities

Roadmap

• PDF content extraction for auto-generation
• Web scraping tools for online content
• Advanced AI card improvements
• Obsidian integration
• More export formats (Notion, Roam)

Show Your Support

If you find Anki MCP useful:

• Star this repository on GitHub
• Share it on social media
• Write a blog post or tutorial
• Create a demo video
• Join discussions and help others

License

MIT License - see LICENSE for details

---

Built for the Anki and MCP communities

Back to top