================== /// MCP /// /// OPE /// ================== [server:online] [protocol:ready]
OpenDataMCP
by OpenDataMCP
Connect any Open Data to any LLM with Model Context Protocol (MCP); includes CLI to install/use MCP servers and a framework to build & publish new open-data providers.
126
17
Open Source01
uvx odmcp
Display all available commands in the Open Data MCP CLI tool
02
uvx odmcp list
List all available Open Data MCP data providers
03
uvx odmcp info $PROVIDER_NAME
Show detailed information about a specific data provider
04
uvx odmcp setup $PROVIDER_NAME
Install and configure the chosen provider’s MCP server for use in the Claude Desktop app
05
uvx odmcp remove $PROVIDER_NAME
Uninstall the chosen provider’s MCP server from the Claude Desktop app
Installation
1. Prerequisites
• Python 3.9 or newer
• git, curl (for testing endpoints)
2. Clone the repository
git clone https://github.com/OpenDataMCP/OpenDataMCP.git
cd OpenDataMCP
3. Create a virtual environment (recommended)
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
4. Install server and core dependencies
pip install -U pip wheel
pip install -e .[server] # installs OpenDataMCP in editable mode with server extras
5. (Optional) Install connector extras if you intend to use them
pip install "OpenDataMCP[postgres,elastic,aws]"
6. Environment configuration (examples)
export MCP_HOST=0.0.0.0
export MCP_PORT=8080
export MCP_ALLOWED_ORIGINS="*" # CORS
# Add datasource-specific variables, e.g. PG_CONN, ELASTIC_URL, AWS_ACCESS_KEY, etc.
7. Launch the service
python -m opendatamcp.server # FastAPI / Uvicorn entry-point
# or
uvicorn opendatamcp.server:app --host "$MCP_HOST" --port "$MCP_PORT"
8. Verify
curl http://localhost:8080/health ➜ {"status":"ok"}
curl http://localhost:8080/docs ➜ interactive Swagger UI
9. Docker (alternative)
docker build -t opendatamcp .
docker run -p 8080:8080 -e MCP_* opendatamcp
Documentation
License: MIT License
Updated 7/30/2025