Apiverket MCP Server

An MCP (Model Context Protocol) server that gives AI assistants access to 120+ Swedish public data endpoints via the Apiverket API. Works with Claude Code, Claude Desktop, Cursor, VS Code, and any MCP-compatible client.

What it does

Two tools, one goal: let AI assistants query Swedish public data using natural language.

Ask "What's the weather in Stockholm?" or "Show train delays from Gothenburg" — the AI finds the right endpoint, fills in parameters, and returns the data. Use govdata_discover first instead of guessing /v1 paths.

Categories

Business & Companies, Culture & Media, Demographics, Economy & Finance, Education, Environment & Nature, Geography, Government & Safety, Health & Medicine, Jobs & Labor, Social Insurance, Telecom, Tourism, Transport & Infrastructure, Weather & Climate

Quick start

Claude Code / Claude Desktop

Add to your MCP config (~/.claude/claude_code_config.json or Claude Desktop settings):

{
  "mcpServers": {
    "apiverket": {
      "command": "node",
      "args": ["/path/to/apiverket-mcp/dist/index.js"],
      "env": {
        "GOVDATA_API_URL": "https://apiverket.se",
        "GOVDATA_API_KEY": "sk_test_demo"
      }
    }
  }
}

Cursor / VS Code

Add to .cursor/mcp.json or .vscode/mcp.json:

{
  "servers": {
    "apiverket": {
      "command": "node",
      "args": ["/path/to/apiverket-mcp/dist/index.js"],
      "env": {
        "GOVDATA_API_URL": "https://apiverket.se",
        "GOVDATA_API_KEY": "sk_test_demo"
      }
    }
  }
}

npx (no install)

{
  "mcpServers": {
    "apiverket": {
      "command": "npx",
      "args": ["-y", "apiverket-mcp-server"],
      "env": {
        "GOVDATA_API_URL": "https://apiverket.se",
        "GOVDATA_API_KEY": "sk_test_demo"
      }
    }
  }
}

Build from source

git clone https://github.com/vinvuk/apiverket-mcp.git
cd apiverket-mcp
npm install
npm run build

Configuration

The test key (sk_test_demo) returns sandbox data for all endpoints — no signup required.

Tools

govdata_discover

Search and browse available endpoints. This is the first tool agents should call when they do not already know the exact Apiverket path.

# List all categories
govdata_discover()

# Search by keyword
govdata_discover(query: "electricity prices")

# Browse a category
govdata_discover(category: "Weather & Climate")

# Search within a category
govdata_discover(query: "forecast", category: "Weather")

govdata_query

Call any endpoint with parameters.

# Current weather
govdata_query(endpoint: "/v1/weather/{city}", path_params: {city: "stockholm"})

# Job search
govdata_query(endpoint: "/v1/jobs/search", query_params: {q: "developer", limit: 5})

# Exchange rates
govdata_query(endpoint: "/v1/rates")

# Train departures
govdata_query(endpoint: "/v1/transport/trains/{station}", path_params: {station: "Cst"})

# Electricity prices by area
govdata_query(endpoint: "/v1/energy/electricity/{area}", path_params: {area: "SE3"})

# Company lookup
govdata_query(endpoint: "/v1/companies/{orgNumber}", path_params: {orgNumber: "5568710426"})

# Police events
govdata_query(endpoint: "/v1/police/events", query_params: {limit: 10})

# VAT validation
govdata_query(endpoint: "/v1/vat/{vatNumber}", path_params: {vatNumber: "SE556871042601"})

How it works

1. The MCP server runs as a local process using stdio transport
2. Your AI clien

... View full README on GitHub