Japanese Anki MCP Server

A Model Context Protocol (MCP) server for Japanese language learning with Anki. Features include:

• Vocabulary Card Creation - Add Japanese vocabulary cards with furigana, readings, and audio
• reMarkable Integration - Extract vocabulary from handwritten notes using Claude Vision OCR
• Bulk Import - Add multiple cards at once
• Full Anki Management - Decks, models, search, tags, sync

Installation

npm install
npm run build

Configuration

Create a .env file (see .env.example):

# Required for reMarkable OCR
ANTHROPIC_API_KEY=sk-ant-...

# Optional: reMarkable cloud access
REMARKABLE_DEVICE_TOKEN=...

Usage with Claude Desktop

Add to your claude_desktop_config.json:

{
  "mcpServers": {
    "japanese-anki": {
      "command": "node",
      "args": ["C:/Users/jnguyen/Documents/Project/MCP/japanese_anki_mcp/dist/index.js"]
    }
  }
}

Available Tools

Anki Connection

• anki_check_connection - Verify AnkiConnect is running

Deck Management

• anki_list_decks - List all decks
• anki_create_deck - Create a new deck
• anki_get_deck_stats - Get deck statistics

Note Models

• anki_list_models - List note models
• anki_get_model_fields - Get fields for a model
• anki_create_japanese_model - Create a Japanese vocabulary model

Notes

• anki_add_japanese_vocab - Add a single vocabulary card
• anki_add_bulk_vocab - Add multiple cards
• anki_search_notes - Search notes
• anki_get_note_info - Get note details
• anki_update_note - Update a note
• anki_delete_notes - Delete notes
• anki_manage_tags - Add/remove tags

Sync

• anki_sync - Sync with AnkiWeb

reMarkable Integration

• remarkable_register - Register with reMarkable cloud
• remarkable_list_notebooks - List notebooks
• remarkable_extract_vocab - Extract vocabulary from an image
• remarkable_extract_vocab_base64 - Extract from base64 image
• remarkable_to_anki - Full pipeline: image → OCR → Anki cards

reMarkable Workflow

Option 1: Cloud Integration

1. Run remarkable_register with a one-time code from https://my.remarkable.com/device/desktop/connect
2. Save the device token to your .env file
3. Use remarkable_list_notebooks to see your notebooks

Option 2: Export Images (Recommended)

1. Export pages from reMarkable as PNG/PDF
2. Use remarkable_extract_vocab with the image path
3. Or use remarkable_to_anki for the full pipeline

Example Usage

# Add a vocabulary card
anki_add_japanese_vocab(
  deck_name="Japanese::Vocabulary",
  model_name="Basic",
  word="食べる",
  reading="たべる",
  meaning="to eat"
)

# Extract vocabulary from handwritten notes
remarkable_to_anki(
  image_path="C:/Users/jnguyen/notes/lesson1.png",
  deck_name="Japanese::Vocabulary"
)

Requirements

• Node.js 18+
• Anki with AnkiConnect plugin (addon code: 2055492159)
• Anthropic API key (for Vision OCR)

License

MIT