Free
$0 / always
The whole platform. Every metric, every integration.
Create an account- All integrations
- Per-activity analysis
- Power-duration and W′ balance
- Time in zone
- Training calendar
- Wellness and readiness
Know who needs you today.
Twenty athletes uploaded something yesterday and you’ve got one morning. Nobody opens twenty files. QuikShift reads each session as it lands, scores load and recovery against that athlete’s own history, and puts whoever’s drifting at the top of the list. Fifteen analyses sit under every name, an AI that’ll explain any of them, and a source under every number. It does the reading. You do the coaching.
$50 a month, ten athletes or a hundred. No per-head fee, and nothing for them to pay.
Pulls from whatever they already train on
- Strava
- Garmin
- intervals.icu
- TrainingPeaks
- Wahoo
- COROS
- Polar
- Suunto
- Zwift
- Whoop
- Oura
- Apple Health
- Google Fit
- Fitbit
- Peloton
- Hammerhead
- Stryd
- Concept2
- and more each month
Four reads on every athlete, every day
Four measurements, each sitting next to the range it’s normal in for that athlete. Showing you the range is the whole point: it’s what lets you look at one and decide the number’s wrong.
- Readiness
- HRV, resting heart rate, sleep, and what the athlete says they’ve got that morning, weighed against their own baseline rather than a population average.
- Ramp and ACWR
- How steeply load is climbing against the last four weeks of it. When a build starts outrunning the base you see it while there’s still room to ease off.
- Fitness and form
- The Banister model fitted per athlete, so the fitness and fatigue constants come out of their own history instead of a textbook default.
- Durability
- Whether power and pace hold through the back half of a long session or quietly fall away. Races turn on this and a session average buries it completely.
Then go as deep as the question needs
Open an athlete and the board’s already built: fifteen analyses fitted to their data, each carrying its slope, its fit, its sample size and the paper the method came from. When the relationship isn’t there, the panel tells you. Efficiency factor on this athlete reports p = 0.109 and calls itself stable, because a null result is still a result.
- Training load
- Fitness trend (eFTP)
- Dose-response
- Efficiency factor
- Form & ramp rate
- Volume & consistency
- Power variability
- eFTP projection
- Anomaly watch
- Signal scan
- Ramp rate & ACWR
- Form (TSB) trend
- Fitness-form readiness
- Monotony & strain
- Fitness momentum
AI that does the reading, not the coaching
It won’t coach for you. What it does is get you to the evidence faster, so the call you make at the end of it is better informed than it would’ve been after opening files one at a time.
- Explains any graph
- Drop the Brain on a chart and it reads that chart for that athlete: what the slope means for them, and whether the fit’s strong enough to act on yet.
- Builds the graph you asked for
- Describe a relationship you want to see and it assembles that view out of the athlete’s data, instead of making you hunt for it among presets.
- Answers questions in context
- Ask about an athlete and the whole computed analysis is already in front of it. You never paste numbers in or explain again who you mean.
The AI only gets to describe what was measured
Most AI training tools answer from whatever they read on the open web. This one has no such licence. The analysis engine computes the numbers first, using the methods on the science page, and the model only gets to put those numbers into sentences.
Then every figure in the write-up is checked back against the computed result. A sentence carrying a number the engine never produced gets dropped before it reaches you. It won’t plan a season, overrule you, or fill a gap with a guess.
Read how the checking worksNo number without a source
Every metric names the method it comes from, and every method links to the paper. The library holds 55 published sources and was last reviewed in May 2026.
- Coggan & Allen (2010)
- Monod & Scherrer (1965)
- Skiba et al. (2012)
- Banister (TRIMP)
- Seiler (intensity distribution)
- Minetti et al. (2002)
Pricing
One price to coach, whatever the size of your roster. Your athletes never see a bill.
Coach
$50 / per month
The roster read, plus everything in Pro for your own training. Ten athletes or a hundred, the price doesn’t move.
Start coaching- Everything in Pro
- Unlimited athletes, free for them
- Roster sorted by who needs you
- Per-athlete analysis and notes
- Shared training plans
Pro
$12 / per month
For training yourself. Adds the analysis layer to your own data.
Go Pro- Everything in Free
- Session analysis and notes
- A read on today’s session
- Trend and durability narration
- New features first
Training on your own?
The same analysis, pointed at one athlete. Connect a service and read your own sessions the way a coach would read them. Free to start, and the depth’s there when you want it.