LearnFlow MCP

LearnFlow MCP is a TypeScript Model Context Protocol server aimed at students, developers, and hybrid learners who need structured planning, note capture, quiz generation, and session memory.

Current MVP

The server currently provides these tools:

• get_profile
• set_profile
• create_learning_plan
• add_work_item
• list_work_items
• save_note
• summarize_notes
• generate_quiz
• save_session_memory
• get_workspace_overview

All workspace data is stored locally in data/learnflow-store.json.

Run locally

npm install
npm run build
npm run dev

For MCP clients that use the built artifact:

npm run build
npm start

Global install

After publishing to npm, users can install the server globally:

npm install -g learnflow-core-mcp

Then MCP clients can launch it with:

{
  "mcpServers": {
    "learnflow": {
      "command": "learnflow-core-mcp",
      "args": []
    }
  }
}

You can also run it without global install via:

npx learnflow-core-mcp

Product direction

This MVP is designed as the foundation for a production-ready server with:

• local-first persistence
• strict schema validation with Zod
• atomic file writes
• reusable session memory
• study and developer workflow support in one server

Suggested next milestones

1. Add update and delete operations for plans, notes, and work items.
2. Add resources for dashboard-style read access.
3. Introduce SQLite for stronger persistence and concurrency handling.
4. Add authentication and tenant scoping if you make it multi-user.
5. Add observability, rate limiting, and audit logs for production deployment.