SCB

SCB MCP server provides seamless access to Statistics Sweden's PxWebAPI 2.0 data. It exposes over 1,200 Swedish statistical tables across demographics, economy, environment, labor, and education, with data updated monthly or quarterly and population data through November 2025. The server supports natural language-like querying, region code resolution, wildcards, and fuzzy matching to help LLMs and AI chatbots fetch precise statistics. To use it, connect your MCP client to the remote HTTP endpoint provided in the README or deploy a local instance and point your MCP host to the local server. The hosted endpoint at https://scb-mcp.onrender.com/mcp is designed for quick starts and requires no authentication. The server adheres to MCP 2025 standards and supports multiple languages (SV/EN) with region-aware results.

Prerequisites:

• Node.js and npm (for local installation) or Docker (for containerized deployment)
• Basic command-line tools (git, shell)

Option A — Remote URL (no installation):

• Use the hosted MCP endpoint directly. No installation required.

Option B — Local installation (Node.js):

1. Clone the repository

git clone https://github.com/isakskogstad/SCB-MCP.git
cd SCB-MCP

2. Install dependencies and build

npm install
npm run build

3. Run the local server (example on port 3000)

npm run start

4. Configure your MCP client to point to the local server, e.g. http://localhost:3000/mcp

Option C — Docker:

1. Build and run the Docker image

docker build -t scb-mcp .
docker run -p 3000:3000 scb-mcp

2. Access the MCP endpoint at http://localhost:3000/mcp

Notes:

• If you use the local installation, you can customize the port by setting PORT before starting (e.g., PORT=8080 npm run start).
• The remote URL option is the quickest path to try MCP capabilities without installation.

Tips and considerations:

• The MCP endpoint supports CORS and does not require authentication for the remote host, but consult any policy changes if you deploy locally with authentication.
• Language defaults to Swedish (sv); you can request results in English (en) if supported by your client.
• Rate limiting for the remote service is enforced (30 requests per 10 seconds) to ensure stability.
• If you encounter region name mismatches, use the region code or take advantage of fuzzy matching (e.g., Goteborg -> Göteborg).
• When running locally, ensure the port you choose (e.g., 3000 or 8080) is open and not used by other services.
• For production deployments, consider running behind a reverse proxy and enabling TLS termination.