TfL MCP. Plan London journeys and track real-time arrivals.

TfL MCP Server - London Transport Data API

Man, this thing gives your AI client real-time access to every piece of data running through London's transit system. Forget jumping between half a dozen different apps or staring at physical arrival boards that are always wrong. Your agent acts like a dedicated transport analyst and journey planner, handling everything from complex door-to-door trips to checking if the Overground is even running today.

Planning Trips Across London

When you need someone to figure out how to get from point A to point B, your client uses get_journey, which calculates multiple travel options combining walking, cycling, tube, and bus into one full itinerary. It gives total duration and step-by-step instructions for the whole ride. To make sure it knows what modes are even available, it calls get_modes, listing every method of travel in the London network, like tram or underground.

If you're trying to get somewhere specific—whether it's a museum or just a weird corner—you can use get_place_search to locate Points of Interest across London by category or coordinates.

Tracking Movement and Stops

For live updates, your agent uses get_arrivals, which predicts the next arrival time for any specific Tube, bus, DLR, tram, or river vehicle at a known TfL stop point. If you're starting with just a name or an address, you gotta find the actual station ID first; that's where search_stop_point comes in, locating and returning IDs for stops based on common names or street addresses.

For bike commuters, it handles two things: get_bike_points lists all nearby docking stations across London coverage areas, showing exactly how many bikes are available and how many docks are empty. If you only care about one specific station, get_bike_point gives the detailed status and location data for that single Santander Cycles dock.

Checking System Health and Roads

To know if things are messed up, your agent checks multiple systems. It runs get_line_status, which reports the current service status—like Good Service or Minor Delays—for any specified TfL line system-wide. For road travel, it pulls fresh data on closures and disruptions using get_road_disruptions, telling you what caused a closure and which roads are affected.

If you're just checking if a specific artery like the A40 is open for driving, you use get_road_status. For general infrastructure knowledge, get_stop_point_details gives full metadata on one stop point, including its fare zone, accessibility features, and station type. Finally, if someone's driving into London, your client can check vehicle compliance by using get_vehicle_details, confirming if a specific registration number meets ULEZ or Congestion Charge rules.

This server lets you build complex logic right inside your agent. You don't need to manually query for arrivals, then check the line status, and then find the nearest station; all those steps happen automatically from one conversation.