Anki MCP
@arielbk
About Anki MCP
Anki MCP gives you conversational access to your Anki flashcards—quiz yourself, create cards, analyze progress, and manage your collection with natural language.
Config
Add this server to your MCP-compatible client using the configuration below.
{
"mcpServers": {
"anki-mcp": {
"command": "npx",
"args": [
"@arielbk/anki-mcp"
]
}
}
}Tools
25Create, update, delete, search, inspect, and tag Anki flashcards through one high-level note/card workflow. Use this for note lifecycle work and card lookup, not for deck structure, note-type schema changes, or study scheduling. Mutating operations change the local Anki collection immediately; searches are read-only. Choose `operation` first, then provide the operation-specific IDs, query, fields, tags, or pagination inputs described by the schema.
Find due cards and apply review-state actions such as answering, suspending, unsuspending, resetting, and relearning cards. Use this for scheduling and review flow, not for editing note content or reorganizing decks. Several operations mutate card state in Anki immediately; `find_due` and `check_status` are read-only. Select `operation`, then pass either `query`, `cardIds`, or `answers` as required for that action.
Create, delete, list, inspect, move cards between, and configure Anki decks. Use this for deck-level organization, not for editing flashcard fields, note types, or study answers. Create, delete, move, and config operations mutate the collection; list and stats/config reads do not. Set `operation` first and supply the matching deck names, card IDs, target deck, or config ID.
Read study analytics and review history for decks, the collection, or specific cards. Use this when the agent needs reporting or diagnostics, not when it needs to mutate Anki data. This tool is read-only but some scopes can return large payloads, so prefer the narrowest scope and explicit card IDs or timestamps when available. Choose `scope`, then provide the deck, cards, or review start timestamp required by that scope.
List, create, and modify Anki note types, including fields, templates, and CSS styling. Use this for schema/template work, not for individual note content, decks, or review scheduling. All operations except `list` mutate the collection and can affect future rendering or compatibility of notes using that model. Pick `operation` first, then provide the model name plus the exact field, template, CSS, or creation inputs needed.
Run general Anki utility operations: sync, version lookup, package import/export, profile listing, and media storage or retrieval. Use this for cross-cutting maintenance tasks that do not belong to flashcard, deck, model, or study workflows. Import, export, sync, and media writes have external side effects; version and profile reads do not. Select `operation`, then provide the relevant deck name, file path, filename, media payload, URL, or pattern.
Retrieve one named Anki media file as base64 data and, for images, as MCP image content that compatible clients can inspect. Use this when an agent needs the actual contents of a referenced media asset, not when it only needs media names or storage operations; those belong in `anki_operations`. This tool is read-only and rejects path traversal. Pass a simple filename exactly as stored in Anki media.
Overview
What is Anki MCP?
Anki MCP is a Model Context Protocol server that connects Anki flashcard collections to AI assistants like Claude, enabling natural language interaction with study materials.
How to use Anki MCP?
Install the package globally with npm install -g @arielbk/anki-mcp, then configure your MCP client (e.g., Claude Desktop) to run the server with the command anki-mcp. Ensure Anki is running with the AnkiConnect plugin installed. Alternatively, clone the repository for a development setup.
Key features of Anki MCP
- Full Anki integration via all 115 AnkiConnect API methods
- Conversational interface for natural language commands
- Bulk operations for managing thousands of cards
- Real-time study statistics and analytics
- Media support for images, audio, and more
- Complex search queries using Anki’s search syntax
Use cases of Anki MCP
- Interactive quizzing on due or struggling cards
- Automated flashcard creation from PDFs or notes
- Analyzing study patterns and retention rates
- Bulk tagging, moving, and deduplicating cards
- Organizing decks and managing card metadata
FAQ from Anki MCP
What prerequisites are needed?
You need Anki installed, the AnkiConnect plugin (add-on code 2055492159), and Node.js 18+ to run the MCP server.
How do I fix “Failed to connect to AnkiConnect”?
Ensure Anki is running and the AnkiConnect plugin is installed and enabled. The server connects to localhost:8765 by default.
What AnkiConnect version is required?
AnkiConnect version 25.2.25.0 or newer is recommended for full compatibility; most features work with older versions.
Can I use this with other MCP clients?
Yes, Anki MCP works with any MCP-compatible client, such as the MCP Inspector, Cline, or custom clients via stdio transport.
Frequently asked questions
What prerequisites are needed?
You need Anki installed, the AnkiConnect plugin (add-on code `2055492159`), and Node.js 18+ to run the MCP server.
How do I fix “Failed to connect to AnkiConnect”?
Ensure Anki is running and the AnkiConnect plugin is installed and enabled. The server connects to `localhost:8765` by default.
What AnkiConnect version is required?
AnkiConnect version 25.2.25.0 or newer is recommended for full compatibility; most features work with older versions.
Can I use this with other MCP clients?
Yes, Anki MCP works with any MCP-compatible client, such as the MCP Inspector, Cline, or custom clients via stdio transport.
Basic information
More Other MCP servers
Mcp
browsermcpBrowser MCP is a Model Context Provider (MCP) server that allows AI applications to control your browser
XcodeBuildMCP
cameroncookeA Model Context Protocol (MCP) server and CLI that provides tools for agent use when working on iOS and macOS projects.
MaxKB
1Panel-dev🔥 MaxKB is an open-source platform for building enterprise-grade agents. 强大易用的开源企业级智能体平台。
Blender
ahujasidOpen-source MCP to use Blender with any LLM
MCP Go 🚀
mark3labsA Go implementation of the Model Context Protocol (MCP), enabling seamless integration between LLM applications and external data sources and tools.
Comments