JorritPeters/sweden-public-transport-mcp-server

Swedish Public Transport MCP Server

A Model Context Protocol (MCP) server that provides access to Swedish public transport information through the ResRobot API. This server allows AI assistants to retrieve real-time transport data for Swedish public transport to plan your journey from A to B.

Prerequisites

• Python 3.8+
• UV package manager (for local development)
• ResRobot API key from Trafiklab
• Google Cloud SDK (for Cloud Run deployment)

Setup

1. Clone the repo

Clone this repository and navigate to the directory:

git clone <repository-url>
cd sweden-public-transport-mcp-server

2. Environment Variables

Create a .env file in the project root:

# ResRobot API Configuration
resbot_key=your_resrobot_api_key_here

Local Development

Running the Server Locally

1. Install dependencies:

   cd transport
   uv sync
   

2. Run the local MCP server:

   uv run mcp_local.py
   

Claude Desktop Configuration (Local)

Add this to your Claude Desktop configuration file (working version for Claude desktop):

{
  "mcpServers": {
    "transport": {
      "command": "uv",
      "args": [
        "--directory",
        "/Users/jorpet/Documents/git/sweden-public-transport-mcp-server/transport",
        "run",
        "mcp_local.py"
      ]
    }
  }
}

Cloud Deployment

Deploy to Google Cloud Run

1. Deploy to Cloud Run:

   gcloud run deploy mcp-server \
     --no-allow-unauthenticated \
     --region=us-central1 \
     --set-env-vars resbot_key=your_resrobot_api_key_here,local_run=false
     --source .
   

Obs: this is a very unsecure way of setting the api key. Next version will use Google Secret Manager.

2. Set up proxy for local access:

   gcloud run services proxy mcp-server --region=us-central1
   

Claude Desktop Configuration (Cloud Run)

For the cloud-deployed version, add this to your LLM configuration (working version for Claude desktop):

{
  "mcpServers": {
    "transport_cloud_run": {
      "command": "npx",  
      "args": ["-y", "mcp-remote", "http://localhost:8080/mcp"]
    }
  }
}

Testing

Test Cloud Run Deployment

Use the provided test script to verify your Cloud Run deployment:

python test_mcp_cloud_run.py

This script will:

• List available tools from the MCP server
• Test the find_station_id_tool with "Uppsala"
• Test the get_travel_options_tool with example station IDs

Manual Testing

You can also test the tools manually through Claude Desktop once configured:

Get travel options:

What are the travel options from Uppsala to Stockholm?

Available Tools

The MCP server provides two main tools for Swedish public transport information:

find_station_id_tool

Finds station IDs by city or station name using the ResRobot location API.

Parameters:

• city_name (string, required): The name of the city or station to search for

Input Example:

{
  "city_name": "Uppsala"
}

Output Example:

Station ID of Uppsala is 740000005

Usage in Claude:

Find the station ID for "Stockholm"

Claude Response: The tool will return station information for Stockholm, including the main Stockholm Centralstation with ID 740000001.

get_travel_options_tool

Gets travel information between two cities using their station IDs.

Parameters:

• start_city_id (string, required): The station ID of the departure city
• end_city_id (string, required): The station ID of the destination city

Input Example:

{
  "start_city_id": "740000005",
  "end_city_id": "740000001"
}

Output Example:

Travel options from 740000005 to 740000001:
1. Departure: 14:17 → Arrival: 14:55 (Duration: 38 minutes)
2. Departure: 14:47 → Arrival: 15:25 (Duration: 38 minutes)
3. Departure: 15:17 → Arrival: 15:55 (Duration: 38 minutes)

Usage in Claude:

What are the travel options from Uppsala to Stockholm?

Claude Response: The assistant will:

1. First use find_station_id_tool to find station IDs for Uppsala and Stockholm
2. Then use get_travel_options_tool with the found station IDs
3. Present you with formatted departure times, arrival times, and journey durations

Common Station IDs:

• Stockholm Centralstation: 740000001
• Uppsala Centralstation: 740000005
• Göteborg Centralstation: 740000002
• Malmö Centralstation: 740000003