Memora MCP
An interactive flashcard MCP App for Claude Desktop. Claude generates flashcards from your request or the conversation and renders them as an inline, gradeable flip-card review with spaced repetition.
README
Memora MCP
An interactive flashcard MCP App for Claude Desktop. Claude generates flashcards from your request or the conversation and renders them as an inline, gradeable flip-card review with spaced repetition. Your grades flow back to the model, so Claude can react to your score and offer to drill what you missed.
Keywords: Model Context Protocol, MCP server, MCP Apps, Claude Desktop, flashcards, spaced repetition, SRS, Anki alternative.
Built on the MCP Apps extension (SEP-1865): core MCP spec 2025-11-25 plus the Apps extension 2026-01-26.
What it does
review_deck: opens a deck as an inline flip-card UI. Click to flip, grade Got it / Missed it, see a results screen. Due cards come first, and grades are sent back to the model via the host bridge.create_deck: Claude generates cards from your request or the chat, saves them todata/decks.json, and renders them immediately. They persist, soreview_deckcan replay them later. Generation follows Memora's card-quality rules (atomic single-concept cards, 1-5 word answers, active-recall or cloze prompts, unambiguous).create_quiz: Claude generates multiple-choice questions (question+options+answer); the UI shows the options, reveals correct/wrong on pick, and schedules them with the same FSRS. Quiz and flashcard decks share the same tree and review flow, and quiz decks are flagged with a badge in the tree and mind map.grade_card(spaced repetition): each grade persists a per-card FSRS schedule (viats-fsrs), so weak cards resurface sooner.due_todaysummarizes what is due across all your decks.- Decks & categories: switch decks from a dropdown, browse a category tree (deck names use
::to nest, e.g.LLM::Attention), or see the same tree as an inline mind map. Thestudytool reviews a whole subtree at once, merging its decks into one session. - Decks are plain JSON read live on every call, so you can hand-edit them or let Claude create them. No database, no external service.
How it works (MCP Apps)
A tool declares a ui:// resource. When Claude calls the tool, the host (Claude Desktop) fetches that resource and renders its HTML in a sandboxed iframe, passes the tool result to the UI, and the UI talks back to the host over JSON-RPC.
Tech stack
- Server: TypeScript,
@modelcontextprotocol/sdkplus@modelcontextprotocol/ext-apps, stdio transport. - UI: React plus Vite, bundled to a single inlined HTML file via
vite-plugin-singlefile. - Runtime is plain
node(no bun or tsx needed) once built.
Prerequisites
- Node.js 20+
Setup
npm install
npm run build # builds the UI bundle (dist/mcp-app.html) and compiles the server (dist/)
Connect to Claude Desktop
Open Settings > Developer > Edit Config and add (use the absolute path to this folder):
{
"mcpServers": {
"memora": {
"command": "node",
"args": ["C:\\path\\to\\memora-mcp\\dist\\main.js", "--stdio"]
}
}
}
Then fully quit Claude Desktop (from the system tray) and relaunch. memora should appear under Settings > Developer.
Usage (in a Claude Desktop chat)
review my World Capitals deckmake me a deck of 10 Spanish travel phrasesturn what we just discussed into a deck called "Photosynthesis"add 5 harder capitals to my World Capitals deck(usesappend)what's due today?
Development
npm run dev # vite watch (UI) plus tsx server on http://localhost:3001/mcp
npm run typecheck # tsc --noEmit (UI)
For fast local iteration you can also run the app against the MCP Apps reference host (basic-host) from the ext-apps repo.
Deck format (data/decks.json)
{
"Deck Name": [
{ "front": "question", "back": "answer" }
]
}
Read live (mtime-cached). create_deck and grade_card write here atomically. Cards gain optional spaced-repetition fields (due, interval, ease, reps) as you review them; cards without them are treated as new. Keep it valid JSON, or the server falls back to a built-in default deck.
Project structure
memora-mcp/
├── server.ts # MCP server: review_deck + create_deck + grade_card + ui:// resource
├── main.ts # entry: stdio (Claude Desktop) or Streamable HTTP transport
├── mcp-app.html # UI entry HTML (bundled by vite)
├── src/
│ ├── mcp-app.tsx # React flip-card UI, deck picker, grade -> model bridge
│ ├── mcp-app.module.css
│ └── global.css # host theme variable fallbacks (light/dark)
├── data/decks.json # editable decks, read live
├── media/flip-card.svg # README image
├── vite.config.ts # single-file bundle config
└── tsconfig*.json
Roadmap
See TODO.md for the full backlog: inline card editing, deck management tools, a fuller FSRS scheduler, npm + MCP Registry publishing, cross-client hosting, and a real demo GIF.
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。