anki-mcp
MCP server for Anki via AnkiConnect, enabling creation, search, and management of flashcards directly from Claude Code or any MCP client.
README
anki-mcp
MCP server for Anki via AnkiConnect — create, search, and manage flashcards directly from Claude Code or any MCP client.
Features
- Card CRUD — create (bulk with duplicate detection), search, update, delete
- Deck management — list, create (supports
::hierarchy) - Archive & sync — export/import
.apkgfiles, trigger AnkiWeb sync - Progress analytics — deck maturity, Interest Heat Score (which topics you retain best), study streak
- Conversation extraction —
/anki extractscans a conversation for learnable moments and proposes cards - Daily progress script — cron-friendly report generator that reads Anki's SQLite directly
- Claude Code skill — interactive card creation via
/ankislash command
Requirements
- Python 3.13+
- uv
- Anki desktop running
- AnkiConnect addon installed (addon code
2055492159) - Claude Code or any MCP client
Installation
git clone https://github.com/chaosisnotrandomitisrhythmic/anki-mcp.git
cd anki-mcp
uv sync
Usage
Add to your MCP config (~/.claude/settings.json, project .mcp.json, or Claude Desktop config):
{
"mcpServers": {
"anki": {
"command": "uv",
"args": ["--directory", "/path/to/anki-mcp", "run", "anki-mcp"]
}
}
}
Tools
| Tool | Description |
|---|---|
add_notes |
Create flashcards (bulk, with duplicate detection) |
search_notes |
Find cards using Anki query syntax |
update_note |
Edit a card's fields or tags |
delete_notes |
Remove cards by ID |
list_decks |
All decks with card counts (new/learn/review) |
create_deck |
Create a deck (supports :: hierarchy) |
export_deck |
Export deck to .apkg (optionally archive + delete) |
import_deck |
Restore .apkg from archive |
sync |
Trigger AnkiWeb sync |
progress |
Learning report with Interest Heat Score |
health |
Check AnkiConnect connectivity |
Configuration
All configuration is via environment variables with sensible defaults:
| Variable | Description | Default |
|---|---|---|
ANKI_MCP_ARCHIVE_DIR |
Directory for .apkg exports/imports |
~/.local/share/anki-mcp/archive |
ANKI_DB_PATH |
Path to Anki's collection.anki2 |
Platform-aware (macOS: ~/Library/Application Support/Anki2/User 1/collection.anki2, Linux: ~/.local/share/Anki2/User 1/collection.anki2) |
ANKI_MCP_OUTPUT_DIR |
Directory for daily progress reports | ~/Anki Progress |
Daily Progress Script
A standalone script that reads Anki's SQLite database and generates a Markdown progress report. Works even when Anki is running (read-only via WAL mode).
# Run manually
cd /path/to/anki-mcp && uv run python scripts/daily_progress.py
# Cron example (daily at 7:03 AM)
3 7 * * * cd /path/to/anki-mcp && uv run python scripts/daily_progress.py
The report includes deck overview, Interest Heat Score, and study streak.
Claude Code Skill
The skills/SKILL.md file provides a Claude Code skill for interactive card creation. It supports:
/anki— manual card creation/anki extract— scan conversation for extractable knowledge and create cards/anki progress— show learning progress
Uninstallation
- Remove the
ankientry frommcpServersin your MCP config - Optionally remove the skill:
rm -rf ~/.claude/skills/anki/ - Optionally delete the archive directory (
~/.local/share/anki-mcp/by default) - Remove the repo:
rm -rf /path/to/anki-mcp
Your Anki cards are stored in Anki itself — nothing is lost by removing this server.
Contributing
See CONTRIBUTING.md.
Security
See SECURITY.md.
License
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。