v0.1.2

Fix/publish workflow (#8)

* name update

* remove pypi environment gate from publish workflow

* restore environment: pypi in publish workflow

* fixing cross platform issues

* version nump

* numpy version fix

* numpy version fix

v0.1.1

v0.1.1 — Bug Fixes & Cross-Platform Hardening

Bug Fixes

• Fixed NoneType crashes in telemetry tools — 8 methods in TelemetryService (get_rpm_trace, get_drs_usage, get_telemetry_segment, get_corner_telemetry, find_max_speed,
  find_min_speed_corner, throttle_application_analysis, braking_point_analysis) would crash with TypeError when telemetry data was unavailable
• Fixed NoneType crashes in corner speed analysis — corner_entry_speed, corner_apex_speed, corner_exit_speed crashed when the requested corner number didn't exist;
  corner_speed_comparison performed arithmetic on None values
• Fixed NaN-to-number crashes — get_driver_points and get_fastest_lap could crash with ValueError when casting NaN floats to float()/int()
• Fixed visualization tools crashing on missing lap data — generate_tyre_degradation_plot, generate_sector_performance_chart, and generate_race_progression_chart now return an
  error dict instead of crashing when s.laps is None
• Fixed pandas deprecation warnings — Replaced deprecated is_timedelta64_dtype, is_float_dtype, is_datetime64_any_dtype calls in serialization with dtype.kind checks,
  preventing future breakage with pandas 3.0

Cross-Platform Improvements

• Fixed cache directory for pip installs — Cache was using ./cache (relative to CWD), which breaks when running f1-mcp from arbitrary directories. Now defaults to
  ~/.f1-mcp/cache consistently across all code paths. Override with F1_CACHE_DIR env var
• Fixed plot output directory — Plots were written to a relative plots/ folder. Now defaults to ~/.f1-mcp/plots. Override with F1_PLOT_DIR env var
• Added headless matplotlib backend — Set matplotlib.use("Agg") so visualization tools work on headless servers, Docker containers, and CI environments without a display

Performance

• Fixed list_seasons performance — Was making 78+ network calls iterating from 1950 to present. Now returns the supported range (2018+) instantly

CI/CD

• Fixed CI cache paths — Test and publish workflows now cache ~/.f1-mcp/cache to match the updated default directory
• Version bump — 0.1.0 → 0.1.1

v0.1.0

---

🏎️ F1 Analytics MCP Server — v0.1.0

Initial release of mcp-f1, a Model Context Protocol server that brings 118+ Formula 1 analytics tools to Claude Desktop, Claude Code, and any MCP client.

What's included

13 tool categories, 118 tools:

• Session & Driver — race results, standings, circuit details, weather, flag events
• Lap & Sector Analysis — lap times, sector deltas, consistency scores, clean lap filtering
• Telemetry — speed, throttle, brake, gear, RPM traces with corner-level analysis
• Strategy — tyre degradation, stint analysis, compound usage, pit windows
• Pit Stops — pit times, pit lane loss, stop summaries
• Race Position — overtakes, battles, gap evolution, lead changes
• Strategy Intelligence — undercut/overcut simulation, optimal pit window, strategy replay
• Telemetry Intelligence — braking points, corner speeds, dirty air, energy deployment
• Advanced Analytics — driver style clustering, aggression index, consistency scoring
• Predictive AI — race winner prediction, overtake probability, tyre cliff, safety car probability
• Visualizations — speed maps, race progression charts, tyre degradation plots, track dominance maps

Installation

pip install mcp-f1

MCP Configuration

{
"mcpServers": {
"f1": {
"command": "f1-mcp"
}
}
}

Powered by

• FastF1 for telemetry and session data
• Ergast API for historical race data
• fastmcp for MCP server framework

---