
F1 · AI voice commentary
A carbon-dark F1 cockpit for the calendar: live maps, real-time session timing, AI voice commentary you can hear and read as you watch, plus archived rounds—all in one Next.js front end. Commentary text comes from language models; spoken output uses text-to-speech—both driven by the same timing as the UI so lines stay faithful to the track.
Jump straight into the app—calendar, circuit, leaderboard, and AI voice commentary—no need to wait for lights out to explore how it behaves.
Coming soon: live AI voice commentary on race day (real sessions). Explore still works now—maps, timing, and the commentary pipeline on the Live Demo Sprint (Chinese Grand Prix 2026 demo data)—no need to wait for lights out.
Explore wires timing and UI state to a commentary service so generated lines and audio stay in lockstep with Formula 1 on track—not after it.
Core · Commentary pipeline
Commentary over WebSockets: language models draft lines from laps, gaps, and timing; text-to-speech turns them into audio when enabled—streamed from an AWS Serverless pipeline the explore view consumes end to end. Live race-day voice commentary during real sessions is coming soon; try the flow on the Live Demo Sprint—it uses Chinese Grand Prix 2026 circuit and session-shaped data so you can explore without race day.
When the lights go out, live session timing and feeds keep the circuit, standings, and AI narration aligned with the real race—not generic filler.
Browse the calendar, pick a round, and jump between completed Grands Prix and upcoming sessions while commentary hooks follow the selection.
SVG layouts with animated progress and tire reads so both you and the model share the same spatial picture of the race.
Standings plus air and track context the narration layer can reference for strategy and atmosphere.
Archive-friendly panels when the chequered flag has fallen, while the commentary service can still summarize the story of the day.
AI + app stack
Backend: AWS Serverless (APIs, streaming, event-driven workers) · OpenAI APIs (language models for commentary, speech synthesis for audio) · WebSocket commentary service
Frontend: Next.js App Router · React 19 · Tailwind CSS v4 · TypeScript · OpenF1 live timing