Skip to main content
Glama
italoo97

study-mcp

by italoo97
OverviewSchemaRelated ServersScoreDiscussions
Python
Local

📚 study-mcp

A Model Context Protocol (MCP) server that turns any study material into a searchable, AI-powered knowledge base — directly inside Claude Desktop.

CI Python License

✨ What it does

Upload a PDF, slide deck, DOCX, image, or HTML file and instantly:

  • 🔍 Search semantically — ask questions in natural language across all your materials

  • 🧠 Generate summaries — get structured summaries saved directly to Notion

  • 🃏 Create flashcards — auto-generate Q&A cards saved to Notion

  • 📂 Manage materials — list and organize everything you've indexed

Related MCP server: Personal Research Assistant MCP

🏗️ Architecture

study-mcp/
├── src/study_mcp/
│   ├── core/
│   │   ├── config.py       # All settings via environment variables
│   │   ├── embeddings.py   # HuggingFace Transformers (local, no API cost)
│   │   └── chunker.py      # Markdown-aware smart text chunker
│   ├── db/
│   │   ├── chroma.py       # Local vector store (ChromaDB)
│   │   └── pgvector.py     # Cloud vector store (Supabase / Postgres)
│   ├── tools/
│   │   ├── ingest.py       # File ingestion via Docling
│   │   ├── search.py       # Semantic search
│   │   ├── notion.py       # Notion integration
│   │   ├── list_materials.py
│   │   └── help.py         # Tools registry
│   └── server.py           # FastMCP server entry point
├── tests/
└── docs/
    └── claude_desktop_config.json

🛠️ Available Tools

Tool

Description

ingest_file_tool

Convert and index any file (PDF, PPTX, DOCX, HTML, images)

search_tool

Semantic search across all indexed materials

list_materials_tool

List all indexed materials

save_summary_tool

Save a summary to Notion

save_flashcards_tool

Save flashcards to Notion

help_tool

List all tools with descriptions and usage examples

🚀 Quick Start

1. Clone and install

git clone https://github.com/italoo97/study-mcp.git
cd study-mcp
poetry install

2. Configure Claude Desktop

Edit ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "study-mcp": {
      "command": "poetry",
      "args": [
        "--directory", "/absolute/path/to/study-mcp",
        "run", "python", "-m", "study_mcp.server"
      ],
      "env": {
        "EMBEDDING_MODEL": "intfloat/multilingual-e5-small",
        "EMBEDDING_DIM": "384",
        "CHROMA_PATH": "./chroma_db",
        "CHROMA_COLLECTION": "study_chunks",
        "CHUNK_SIZE": "512",
        "CHUNK_OVERLAP": "64",
        "NOTION_TOKEN": "secret_...",
        "NOTION_DATABASE_ID": "your-database-id",
        "DATABASE_URL": ""
      }
    }
  }
}

Vector backend is auto-detected:

  • DATABASE_URL empty → uses ChromaDB locally (zero setup)

  • DATABASE_URL set → uses pgvector on Supabase (free tier works)

3. Restart Claude Desktop

The tools will appear automatically in Claude Desktop.

🧠 Embedding Models

Set EMBEDDING_MODEL to any HuggingFace feature-extraction model:

Model

Languages

Dims

intfloat/multilingual-e5-small

PT + EN + more

384

BAAI/bge-small-en-v1.5

EN only

384

neuralmind/bert-base-portuguese-cased

PT only

768

Update EMBEDDING_DIM if you change models.

📋 Notion Setup

  1. Go to notion.so/my-integrations and create an integration

  2. Create a database with these properties:

    • Name → Title

    • Type → Select (Summary, Flashcards)

    • Material → Rich text

    • Content → Rich text

    • Tags → Multi-select

  3. Share the database with your integration

  4. Set NOTION_TOKEN and NOTION_DATABASE_ID in the config

💡 Example Usage in Claude

"Ingest this file: /Users/me/Downloads/algorithms_lecture.pdf"

"Search my materials for: what is dynamic programming?"

"Generate 10 flashcards from material abc-123 and save to Notion"

"Summarize material abc-123 and save to Notion with tags: algorithms, CS"

"List all materials I've ingested"

🔧 Development

# Install dependencies
poetry install

# Run lint
poetry run task lint

# Run type check
poetry run task type_check

# Run tests
poetry run task test

# Format code
poetry run task format

📄 License

MIT

This server cannot be installed

F
license - not found
-
quality - not tested
B
maintenance

How are these scores calculated?

Maintenance

–Maintainers
–Response time
–Release cycle
–Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/italoo97/study-mcp'

If you have feedback or need assistance with the MCP directory API, please join our Discord server