Forensic analytics for serious triathletes
Built for athletes who'd rather see numbers than vibes.
Off-the-shelf training apps tell you whether you ran today. tri-ai tells you whether your tempo intervals held against your prescription, whether your bike power dropped in the second hour, whether your run pace drifted in your last brick, and what to adjust next time.
Behind the scenes it pulls every data point your watch already captured, layers in established sports-science models (CTL / ATL / TSB, fuel-per-hour, off-bike pace penalty, brick-derived calibration), and produces lap-by-lap analytics you'd otherwise need a spreadsheet, and a coach, to derive.
- Age-group triathletes targeting Olympic, Half, or Full Iron distances.
- Athletes who already have a structured plan (or write their own) and want to see how compliantly they're executing it.
- People who'd rather see numbers than vibes, and who want an AI coach review of every session, including the easy spins and recovery swims a personal coach would never bill for.
- Garmin ConnectLive
Activities, daily wellness (sleep, HRV, body battery, readiness), profile thresholds, planned workouts, body composition (Index scale), personal records.
- TrainingPeaksLive
Activities, daily wellness (sleep stages and quality, HRV, resting heart rate, SpO2, stress, steps), body composition, and your training thresholds, all read through the official TrainingPeaks partner integration. The intended long-term source of truth: TrainingPeaks holds the plan and the calendar, tri-ai reads and interprets them.
- Anthropic ClaudeLive
Powers two AI surfaces: per-workout analyses you generate on demand (what worked, what didn't, what to adjust, with voice configurable across balanced, conservative, aggressive, race-week), and Ask Coach, a conversational assistant that answers questions about your training, always from your own numbers rather than guesswork. Both run on Anthropic's zero-retention tier: prompts aren't retained beyond the request and are never used to train models.
- Open-MeteoLive
Race-day weather forecasts and climate proxies for course conditions.
tri-ai is operated from the United Kingdom by an independent developer with a background in triathlon coaching analytics.
Athlete data is hosted in the EU (London region, eu-west-2), encrypted in transit and at rest. See our Privacy Notice for the full breakdown of how data is handled.
Bug reports, feature ideas, partnership enquiries: hello@tri-ai.app.
Response within 2-3 working days.