Kolada

Provides AI applications with access to Sweden’s municipal KPI data via MCP tools for semantic search, metadata, and cross-municipality analysis.

python

14

GitHub Stars

python

Language

5 months ago

First Indexed

2 months ago

Catalog Refreshed

Documentation & install

Readme and setup notes from the catalogue, plus a client-ready config you can copy for your MCP host.

Installation

Add the following to your MCP client configuration file.

Configuration

View docs

{
  "mcpServers": {
    "aerugo-kolada-mcp": {
      "command": "kolada-mcp",
      "args": [],
      "env": {
        "PORT": "8001",
        "LOG_LEVEL": "INFO",
        "MCP_TRANSPORT": "stdio"
      }
    }
  }
}

Kolada MCP Server is a clean, standalone MCP server that enables AI applications to query Sweden’s municipal and regional KPI data. It provides semantic KPI access, supports multiple tools, and runs without external dependencies, making it practical for AI-assisted analytics and natural language queries over public sector data.

How to use

You interact with the Kolada MCP Server through an MCP client. Run the server in one of its supported modes, then connect your client to issue KPI queries, compare KPIs, fetch data, or perform analyses across municipalities.

Default behavior starts the server in stdio mode, which runs as a local process you invoke directly. You can also start the server in HTTP mode to expose a TCP endpoint on a port you specify. Use the tools in your client to list KPI areas, search KPIs semantically, retrieve KPI metadata, fetch raw data, compare KPIs, and perform cross-municipality analyses.

For Claude Desktop or other suite integrations, point your client at the Kolada MCP endpoint using the appropriate command or transport, then execute queries through your integration layer. The server ships with a set of tools that cover area listing, KPI retrieval, metadata access, data fetching, comparative analyses, and municipality filtering.

How to install

Prerequisites you need on your system: Python 3.11 or newer, and a compatible shell environment.

Option A: Install via pip from the local project directory.

pip install -e .

Option B: Run the server locally using Python directly in stdio mode.

kolada-mcp
# or
python -m kolada_mcp

Option C: Run the server in HTTP mode on a specific port.

MCP_TRANSPORT=http PORT=8001 kolada-mcp

Option D: Start via Docker using the provided container configuration.

docker-compose up -d kolada-mcp

Configuration and usage notes

Environment variables control how the MCP server runs. The following variables are supported for transport, port, and log verbosity.

MCP_TRANSPORT=stdio PORT=8001 LOG_LEVEL=INFO

Tools and capabilities

The server exposes a set of 9 MCP tools to access and analyze the KPI data.

Troubleshooting tips

If the server fails to start, verify that your Python environment is 3.11+ and that you have network access if using a remote transport. Check that the port is not already in use when starting in HTTP mode.

Available tools

list_operating_areas

List all KPI categories with counts.

get_kpis_by_operating_area

Get KPIs within a specific category.

search_kpis

Semantic search for KPIs using natural language.

get_kpi_metadata

Get detailed metadata for a specific KPI.

fetch_kolada_data

Fetch raw KPI data for municipalities.

analyze_kpi_across_municipalities

Comparative analysis with rankings across municipalities.

compare_kpis

Compare two KPIs for difference or correlation.

list_municipalities

List municipalities and regions.

filter_municipalities_by_kpi

Filter municipalities by KPI threshold.