Agmarknet MCP Server

An AI-native Model Context Protocol (MCP) server that provides Large Language Models (like Claude) with real-time access to Indian agricultural commodity prices across 3,000+ markets.

This server acts as a bridge between LLMs and the Indian Government's Agmarknet database (via data.gov.in), allowing you to ask natural language questions like "Where is the cheapest place to buy onions in Maharashtra today?"

Features

• MCP Protocol Native: Built with the official Anthropic mcp SDK using the modern FastMCP architecture.

• Graceful Degradation (Offline Mode): Government APIs frequently block requests via Akamai WAFs or experience downtime. This server implements a robust fallback architecture that automatically switches to a sample offline dataset if the primary API is unreachable. (Note: Until you provide a valid data.gov.in API key, the server uses static sample data for demonstration purposes, which does not reflect live market prices).

• Strict Data Validation: Uses pydantic to enforce rigid data schemas on government API payloads, preventing AI hallucinations caused by malformed data.

Related MCP server: Agricultural AI MCP Server

Available AI Tools

The server exposes the following tools to the LLM:

1. get_commodity_price(commodity, state?, district?, market?): Get daily wholesale prices (min/max/modal) for a specific commodity.

2. compare_markets(commodity, state?, top_n=10): Rank markets to find the absolute cheapest places to buy a specific commodity.

3. get_price_summary(commodity, state?): Get statistical aggregates (average price, spread, most expensive market).

4. list_commodities(search?): List all commodities available in the database today.

5. list_markets(state?, commodity?): List all reporting mandis (markets).

Quickstart

1. Installation

Clone the repository and install the Python package:

git clone https://github.com/Krishna-Baldwa/agmarket-mcp.git
cd agmarknet-mcp

# Create virtual environment and install
python -m venv .venv
source .venv/bin/activate
pip install -e .

2. Configuration

Create a .env file in the root directory:

cp .env.example .env

Add your data.gov.in API key. If you leave it empty or the API is unresponsive, the server will automatically use the built-in Offline Mock Data.

3. Usage with Claude Desktop

To use this with Claude Desktop, add the following to your claude_desktop_config.json file (typically located at ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

{
  "mcpServers": {
    "agmarknet": {
      "command": "/absolute/path/to/agmarknet-mcp/.venv/bin/python",
      "args": [
        "-m",
        "agmarknet_mcp.server"
      ],
      "env": {
        "DATA_GOV_IN_API_KEY": "your_api_key_here"
      }
    }
  }
}

Restart Claude Desktop, and you can immediately start asking:

• "Compare the price of tomatoes across all markets in Gujarat."

• "Give me a statistical summary of Onion prices."

Testing with MCP Inspector

You can interactively test the tools without an LLM:

npx @modelcontextprotocol/inspector python -m agmarknet_mcp.server

Architecture

• server.py: Handles stdio transport and JSON-RPC lifecycle via FastMCP.

• api.py: Manages HTTP traffic via httpx and handles the graceful fallback mechanism.

• models.py: Pydantic models for strict type enforcement.

Data Source

Commodity prices are sourced from the Directorate of Marketing and Inspection (DMI), Ministry of Agriculture and Farmers Welfare, Government of India, via the data.gov.in platform.

License

MIT License