airace listens to your live F1 telemetry over UDP, runs it through an LLM pipeline, and speaks engineer-style radio calls back to you in real time.
Calm, terse, factual — modelled on real F1 race engineers. One snapshot in, one line out. Silent until it has something worth saying. And you can talk back.
There's no hosted service and nothing to pay us. airace is open source and runs entirely on your machine.
Plug in your own API auth. It runs on whatever provider you already pay for — or fully local for zero cost. You pay your provider directly. Nothing flows through us.
Your keys, your models, your data.
Four moving parts, one calm voice in your ear.
Your game streams telemetry over UDP on port 20777. airace catches every snapshot — speed, tyres, fuel, gaps, damage.
Only the data that matters to the moment (or your question) gets sent on. Tyres query → just tyre data. Fewer tokens, faster answers.
A fast local model handles simple questions in seconds and only escalates the hard ones to your main model. Memory keeps it from repeating itself.
A Piper neural voice reads the call out loud, engineer-style. Silent until there's something worth saying.
Your questions never wait on the slow model. Three lanes, each tuned for its job.
Answered by a fast small model — often sub-second. Ask anything and it comes straight back, no waiting on the heavy lifting.
Pre-recorded and played immediately. Critical calls bypass the AI entirely — zero delay when it counts.
Designed to say things still true a lap later, so the round-trip latency simply doesn't matter.
A full race-engineer stack — voice, memory, routing, dashboard — running on your own hardware.
Speaks radio calls out loud via a Piper neural voice. Calm, terse, factual — no coaching fluff.
Voice input via VAD + Whisper speech-to-text. Ask “gap?”, “fuel?”, “should I box?” and it answers.
Safety car, VSC, red flag, blue flag, “box this lap”, SC ending — played immediately, no AI delay.
A fast 3B model handles simple questions in <3s and only escalates hard ones — with a “give us a second” bridge.
Sends only the telemetry that matters to your question — tyres query gets just tyre data. Fewer tokens, faster answers.
Remembers recent calls and 13 topics — gap, tyres, fuel, ERS, weather, damage, pit plan, strategy, lap times — so it never repeats itself.
Speed, gear, RPM, throttle/brake/steer, lap & sector times, tyre temps, and a rolling 60-second trace chart.
Flip a toggle to view the dashboard from your phone or another device on your WiFi.
English, Español, Italiano, Français, Deutsch, Português, Čeština, Polski, 日本語.
OpenRouter, Groq, Anthropic, local Ollama — or any custom OpenAI-compatible endpoint.
Chattiness slider, mute under yellow flag, suppress fuel chatter, tick interval, custom instructions, or override the whole prompt.
Pick any input/output device. Local or cloud (Groq) transcription. First-run wizard; desktop app or headless.
Older games (F1 21–24) are selectable but unverified. Here's exactly where things stand.
The app is on its way — the repo's still private for now. Follow my GitHub profile and subscribe on YouTube to catch it the moment it drops.