# F1InsightsHub

> Formula 1 analysis built on FastF1 timing data: lap times, stints, race pace and distance-aligned telemetry comparisons. The website is https://f1insightshub.com. Agents and AI assistants can use the MCP server below, which returns the same processed data and charts as the website's Analysis tab.

## MCP server

- Endpoint: https://mcp.f1insightshub.com/mcp (MCP Streamable HTTP, stateless, no authentication)
- Claude: Customize > Connectors > + > Add custom connector, paste the endpoint URL, leave OAuth empty.
- ChatGPT (paid plans): Settings > Apps & Connectors > Advanced settings > Developer mode on, then Apps & Connectors > Create, paste the endpoint URL, authentication "No authentication".
- Anything else: any MCP client that speaks Streamable HTTP.

## Tools

- find_sessions: Finds sessions by season, event and session name, and returns their keys
- list_drivers: Lists the drivers in a session
- get_laps: Lap and sector times, tyres, stints and fastest laps, with a lap time chart
- compare_telemetry: Distance-aligned comparison of up to eight laps: corner splits, channels, CSV and charts
- get_telemetry_insights: The site's model of each lap's strengths, weaknesses and missed time
- get_race_pace: Average and robust race pace for the field, with a chart
- get_lap_distribution: Sorted lap time distribution per driver, with a chart
- get_intervals: Gap to the leader and position on every lap, with a chart
- get_session_summary: Best laps, theoretical bests, consistency and team estimates

## How to use

F1InsightsHub (f1insightshub.com) Formula 1 analysis tools. They return the same processed data and charts as the website's Analysis tab.

Typical flow:
1. Identify the session. Every data tool accepts either session_key (from find_sessions) or year + event + session, e.g. {year: 2026, event: "Azerbaijan", session: "Qualifying"}. Event text can be a country, city, circuit or Grand Prix name. Session text can be "Q", "quali", "race", "sprint", "SQ", "FP1" and similar.
2. Drivers can be given by car number, three-letter code (VER), surname or full name. Call list_drivers when unsure.
3. Laps: get_laps lists each driver's laps with times, sectors, tyres and stints. In compare_telemetry a lap is a lap number or "fastest".
4. compare_telemetry is the telemetry comparison. It runs the site's distance-alignment pipeline and returns a per-corner breakdown (segment time, apex speed, braking point, throttle pickup), sampled channels, a full-resolution CSV link and charts. get_telemetry_insights adds the site's model-based strengths and weaknesses per lap.
5. For races and sprints: get_race_pace, get_lap_distribution, get_intervals. get_session_summary covers any session.

Units: distance in metres from the start/finish line along the reference lap, speed in km/h, times in seconds. "gap_to_ref" and "delta_to_ref" are positive when the lap is slower (behind) the reference lap. Chart images come with a public link (valid about an hour) that you can show the user. Data comes from FastF1 timing and telemetry. A session that has not happened yet has no data. The first request for a session can take up to a minute or two while the server loads it.

## Example

"Using F1InsightsHub, compare the fastest qualifying laps of George Russell and Max Verstappen at the 2026 Azerbaijan Grand Prix and tell me where each is stronger."
→ compare_telemetry({"year": 2026, "event": "Azerbaijan", "session": "Qualifying", "laps": [{"driver": "RUS", "lap": "fastest"}, {"driver": "VER", "lap": "fastest"}]})
→ optionally get_telemetry_insights with the same laps.

## Limits

Each client gets about 30 tool calls a minute, and only a few heavy analyses run at once. Chart and CSV links expire after about an hour.