title: Indian Legal MCP emoji: ⚖️ colorFrom: yellow colorTo: gray tags:

• mcp
• legal sdk: gradio sdk_version: 6.27.0 app_file: app.py pinned: false license: mit

⚖️ Indian Legal MCP

Structured, sourced access to 10 major Indian legal texts — built for AI agents.

Live Server • Tools • Acts Covered • Credits • License

📖 What is this?

Indian Legal MCP is a Model Context Protocol server that gives AI assistants — Claude, ChatGPT, or any other MCP-compatible client — exact, structured access to the text of 10 major Indian Acts, instead of relying on the model's memory (and risking hallucination on section numbers, punishments, or repealed law).

Connect it once, and any AI agent using it can pull the real text of a section, browse a chapter, search by keyword, or check whether a law is still in force — all sourced directly from the underlying Act text, not guessed.

Disclaimer: This is not legal advice. Always verify anything that matters against the official Gazette at indiacode.nic.in. Historical Acts (IPC, CrPC, Indian Evidence Act) are clearly marked as repealed in every response — they remain useful for matters predating 1 July 2024, but no longer apply to new cases.

✨ Features

• 10 major Indian Acts in one consistent, queryable structure
• Smart keyword search — exact title matches rank first, partial title matches next, body-text matches last (BM25-style relevance), so the most relevant section always surfaces first
• Safe pagination — large Acts (like BNS's 358 sections) are automatically chunked into manageable pages, so a single query never overflows an AI's context window
• Historical-law awareness — every response involving IPC, CrPC, or the Indian Evidence Act automatically notes that the Act is repealed and names the current replacement (BNS, BNSS, or BSA respectively)
• No embeddings, no ML model running — plain, fast, full-text search under the hood, which keeps this server free to host and quick to respond
• Zero setup for the AI side — connect once, and the AI decides on its own when a question calls for a lookup

📚 Acts covered

🛠️ Available tools

🔌 Connect to the live server

Server URL:

https://gsms-b-indian-legal-mcp.hf.space/gradio_api/mcp/

Claude Desktop

Open Settings → Developer → Edit Config, and add this to claude_desktop_config.json:

{
  "mcpServers": {
    "Indian Legal MCP": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "https://gsms-b-indian-legal-mcp.hf.space/gradio_api/mcp/"
      ]
    }
  }
}

Restart Claude Desktop — the tools will appear under the MCP/tools icon in the chat window.

Other MCP clients (Cursor, Cline, Windsurf, Claude Code, etc.)

For your MCP client’s setup instructions, use the following server URL:

https://gsms-b-indian-legal-mcp.hf.space/gradio_api/mcp/

🖥️ Run it yourself / develop locally

git clone https://github.com/GSMS-B/indian-legal-mcp.git
cd indian-legal-mcp
pip install -r requirements.txt
python normalize.py   # builds legal.db from the raw JSON in data/raw/
python app.py          # starts the Gradio app with MCP enabled

The app will print a local URL — open it in a browser to test each tool by hand, or connect a local MCP client to the printed /gradio_api/mcp/ endpoint.

🙏 Credits & data sources

This project would not have been possible without the excellent datasets compiled by GSMS-B and Civictech India ("For the People & By the people"):

• Indian-Legal-Sections (BNS, BNSS, BSA) — The text for the modern criminal codes (BNS, BNSS, and BSA) was independently compiled and structured for this project by GSMS-B on Hugging Face.
• Indian-Law-Penal-Code-Json — Structured JSON used as the foundational data source for this project's historical Act coverage, including IPC, CRPC, IEA, CPC, MVA, and NIA.
• constitution-of-india — The source dataset used for the Constitution of India (COI).

Huge thanks to Civictech India for making their data freely and openly available. Please check out and star their repositories if you find this project useful.

📄 License

This project is licensed under the MIT License — see the LICENSE file for details. The underlying legal texts themselves are public government documents; the credited source datasets carry their own respective licenses, viewable in their original repositories linked above.