Four things most apps are missing. We run on all four.
Real elite racing. Accepted sports science. Professional coaching. AI. Take any one away and you get a worse product. Here is what each pillar contributes, and what it deliberately does not claim to do.

One engine, four sources of intelligence
Most “AI coaches” start at pillar four and stop there. We start at one, two and three; the AI just delivers it, the expertise is human.
Elite racing experience
From racing at the front
Sports science
Backed by proven studies
Professional coaching
A real coach's method
Artificial intelligence
Delivers it, tailored to you

Elite racing experience
The judgement of someone who has raced at the front: a HYROX Elite athlete with a 55:29 PB and a 4× Spartan World Champion. What a split actually feels like, where races are really decided, and which "good" numbers are quietly costing you the PB.
- In the product
- How the analysis weights the roxzone, the transitions, and the back-half fade: the places most tools never look.
- What it does not claim
- That experience replaces measurement. It is why we measure the right things.

Sports science
Every recommendation is backed by proven exercise-science research and real-world testing: energy systems, pacing, fatigue, specificity, and how training carries over to the eight stations and the runs between them.
- In the product
- Why a plan prioritises one quality over another for your ranked limiters, biggest first, instead of the same template for everyone.
- What it does not claim
- Novel studies or magic protocols. We apply established principles honestly. We do not cite invented research.

Professional coaching
The decision-making of a real coach (rank the limiters biggest first, sequence the work, re-check, adjust) proven across 300+ athletes since 2020 at a 95% success rate. Plus the human option: the same coaching team takes on athletes 1:1.
- In the product
- The weekly plan regeneration and the "what to train next" logic: Richard’s coaching method, made always-on.
- What it does not claim
- That software replaces a coach in the room for everyone. It scales a coach’s judgement; for hands-on accountability, 1:1 exists.
- 01Run fade
- 02Sled push
- 03Wall balls
- 04Roxzone
Limiter: Wall balls under fatigue. Hold Run 6–8 pace to claw back an estimated 2:10.
Artificial intelligence
The AI does not invent the training. It takes the first three pillars, tailors them to your data and your situation, delivers it in seconds, and re-adapts every week as new results and feedback come in. In practice it works like 1:1 coaching, scaled to everyone.
- In the product
- Instant analysis, the AI coach chat, race predictions, and weekly plan regeneration off your real splits.
- What it does not claim
- To be the source of the intelligence. It is the delivery layer, not the expertise. The messenger, not the message.
Five specialised agents, one read of your race
The AI is not a single chatbot guessing. It is a small team of focused agents, each doing the job a coach does by eye, applied to your splits in seconds.
Splits analyst
Reads every run, station and transition from your official results: the raw shape of your race.
Benchmarker
Ranks each split against the real field so a "good" time is judged honestly, not in a vacuum.
Limiter finder
Ranks your limiters biggest first: the biggest leak to target now, then the rest in order.
Predictor
Projects your realistic finish and the time on the table as you fix your limiters, biggest first.
Planner
Builds the plan a coach would build to target the biggest limiter first, then the rest, and regenerates it weekly.
Principles we apply, honestly
Established exercise science, applied to your data. No invented studies, no magic protocols.
- Compromised runningHYROX running happens under fatigue, between stations, so we judge and train it that way, not as fresh-legged pace.
- Aerobic baseA bigger engine carries the back half. Most finish-time gains live in durable aerobic capacity, not a single hero session.
- Pacing under fatigueThe back half is decided by front-half decisions. Pacing is a trainable skill with its own plan and race-day strategy.
- Training specificityYou adapt to what you actually practise. The work targets the eight stations and the compromised runs between them.
- PeriodisationLoad is sequenced and recoverable across your race calendar: sustainable beats heroic, every block.
- Intervals & tempoStructured interval and tempo work builds the top-end and the repeatable race pace HYROX demands, not just easy volume.
- Threshold workTraining at and around threshold raises the pace you can hold before you fall apart, which is exactly what the back half of a race asks for.
- StrengthReal strength work makes the sleds, carries and lunges cost less, so you reach the runs fresher and hold form under fatigue.
- HYROX-specific drillsWe rehearse the actual stations and transitions (wall-ball rhythm, sled mechanics, clean roxzone moves) so race day holds no surprises.
This is the science, rendered from your real data.
Every chart on your dashboard is generated live from your own data, whether that is your official race results or your completed training if you have not raced yet. It is not a screenshot and not a template. The same score below is the one the app draws the moment you import a race or log a session. The sample numbers here are just an example; the rendering is real.


Richard HynekFounder & Head Coach
The intelligence in the engine was earned on the start line and through years of real coaching: a HYROX Elite racer and 4× Spartan World Champion, six years coaching 300+ athletes. The AI just makes it instant, personal and always-on.
- 4× Spartan World Champion
- 3× Spartan European Champion
- Winner of the hardest Spartan Race in history
Your training, rebuilt every week from your real training data.
Start today, backed by our 90-day money-back guarantee*. A personalized plan every week, and re-generate it any time your week changes.