CensusChat

CensusChat provides a natural language interface to US Census data tailored for healthcare strategy teams. It leverages an MCP HTTP transport to accept plain-English queries, translates them into validated SQL, executes them against a DuckDB-backed dataset, and returns interactive results including sortable tables, charts, and export options. You can ask for demographic compositions, market analyses, and healthcare-oriented indicators (e.g., Medicare eligibility, population health indicators) and compare regions side-by-side. The system supports conversational follow-ups like “Show seniors in Tampa Bay” and “Now filter to income over $75K,” preserving context across steps and delivering quick exports to Excel or CSV. Parallel querying and drill-down capabilities enable rapid comparisons across counties and block groups, making it practical for strategy planning and facility placement work.

Prerequisites:

• Docker is installed and running on your machine or deployment environment.
• Access to required API keys: AnthropAI/Claude (Anthropic) and Census data API.
• Optional: a suitable deployment environment for persistent storage and logs.

Step-by-step:

1. Pull and run the CensusChat MCP server container:

# Run the CensusChat MCP server container (exposes its UI/API on localhost:3000 by default)
docker run -it --rm \
  -p 3000:3000 \
  -e ANTHROPIC_API_KEY="<your-anthropic-key>" \
  -e CENSUS_API_KEY="<your-census-key>" \
  -v censuschat_data:/data \
  hollandkevint/censuschat:latest

2. Verify startup and access the UI:

Open http://localhost:3000 in your browser to start querying Census data.

3. For local development without Docker, ensure Node.js-based components and dependencies are installed (if you have a fork or local build). The project uses Docker for deployment in this guide, so the Docker approach is recommended for a straightforward setup.

Tips and considerations:

• Ensure your API keys are valid and not expired before starting the container; the container must be able to reach the external APIs.
• If you need to scale queries, consider using a container orchestration platform and ensure persistent storage for audit logs.
• The MCP validation layer enforces security policies; if you encounter blocked queries, review the allowlists and query patterns.
• For production deployments, set proper TLS termination and rotate API keys regularly. Monitor query latency to maintain sub-2-second response goals.
• If you encounter port conflicts, adjust the -p flag to expose a different host port (e.g., -p 8080:3000).