↗
Predict
See how the training ahead is expected to affect your fitness and move you toward the goal that matters.
YOUR DATA → YOUR PLAN → YOUR LIKELY DIRECTION EARLY BETA · BUILT AROUND THE ATHLETE
Predict.
Adapt.
Motivate.
Helping you run pain-free and achieve your running goals.
Most running tools either record what already happened or tell you what to do next. RunAI is being built to connect the exact training ahead to where it is likely to take you — then recalculate both the plan and the outlook when your body, completed training or life changes.
EARLY BETAStrava connectedGarmin · COROS · Apple Watch direct integrations in progressPricing not yet announced
PREDICTWhere this plan could take you
CONCEPT PREVIEWYOUR GOALBuild toward your next race
On track
TodayWeek 1Week 2Week 3Week 4
PLANNED TRAINING4 weeks ahead
→PREDICTED EFFECTFitness moves toward goal
Illustrative. The product is being built to forecast the likely effect of your actual plan with uncertainty, not promise a result.
THE RUNAI DIFFERENCEMore than a training plan.
More than a training plan.
A learning loop around you.
MOST RUNNING TOOLSPlan → run → analyse what happened
RUNAI'S INTENDED LOOPPlan → predict likely direction → run → learn → adapt plan → update forecast
ONE RUNNER SCENARIO10K PB target. Four weeks are trending the right way. Then a calf niggle and unexpectedly high effort arrive.
Instead of blindly preserving the calendar, the next hard session backs off and the goal outlook updates with the changed plan. Illustrative example — not a guaranteed performance result.
↻
Adapt
Completed training, pain, fatigue, RPE and journal feedback become new evidence for what should happen next.
NEW EVIDENCE → NEW PLAN🏆
Motivate
Celebrate meaningful progress — not just completed checkboxes — with PBs, level upgrades, LT improvements and milestones.
PROGRESS YOU CAN SEECURRENT PRODUCT UI
See the product, not a mock-up.
Captured from the current RunAI app using demo data so no athlete's private training data is published.
01 · CHOOSEChoose the training purpose
The current Add AI Workout flow groups sessions by the training job they are meant to do.
02 · UNDERSTANDSee what the session is for
Workout detail explains the training goal and shows the athlete-relative level and target before it reaches the calendar.
03 · PLANPut the prescription on the plan
The current calendar keeps the planned dose visible with duration, pace, category and RunAI provenance in one compact card.
Real interface, demo data. These screens show current product behaviour; the forecast above remains explicitly experimental.
PREDICT — IN MORE DETAILNot just what to run.
Not just what to run.
What it is likely to change.
RunAI is being built to estimate how the training ahead may affect your fitness and goal trajectory, using your recent training, current capability, completed sessions and the exact work still on the calendar.
The prediction should move when the evidence moves. Change the training, miss a session, improve faster than expected or struggle with the planned dose, and the outlook should be recalculated.
Recent training, current levels, history and what has actually been completed.
Duration, intensity, workout type, progression and how those sessions fit together.
How fitness and goal progress may change if the plan is completed as expected.
Predictions are estimates. Confidence should improve as RunAI learns more about the athlete.
ADAPT
Your plan should adapt when you do.
RunAI should not react to one magic readiness number. It should build a picture from the evidence you create every day — what you ran, how hard it felt, how your body feels and what is going on around the training.
The aim is not to avoid hard training. It is to choose the right hard training, at the right time, for the athlete in front of us.
ILLUSTRATIVE ADAPT DEMOChange the evidence. Watch the plan respond.
Not training adviceRecent sessions feel appropriate and no new niggle is reported.
3 × 8 min controlled threshold effort stays on the calendar.
The illustrative outlook keeps the same direction because the evidence has not changed.
This interaction demonstrates RunAI's intended evidence → plan → outlook loop. The scenario and response are illustrative, not a validated prediction or medical recommendation.
Pain & niggles
Where it hurts, how severe it is and whether running changes it.
Fatigue
Recent load and how recovered you actually feel — not just what the calendar expected.
RPE
A session that felt much harder or easier than prescribed is new evidence.
Journal
Sleep, stress, soreness, illness and the context only you can report.
Completed training
What you really did matters more than what was originally planned.
Missed or changed sessions
Life happens. The rest of the plan should respond instead of pretending it did not.
MOTIVATE
Your hard work deserves more than a checkbox.
RunAI's Trophy Case is designed to recognise progress that normal running apps often miss. A faster interval is good. Repeating it for the first time is better evidence. A level upgrade, a new LT pace or a pain-free streak should feel like something happened.
The trophy is the celebration. The underlying training evidence is what matters.
TROPHY CASEReal progress. Remembered.
6 recent🏆
Interval PB2 × 1 km @ 3:50/kmFirst time repeating the pace
↗
Threshold level up4.7 → 4.9Capability moved forward
⚡
LT pace improved4:34 → 4:05/kmPrediction converted into progress
🔥
Consistency4 weeksTraining kept moving
♡
Pain-free streak6 weeksProgress without forcing it
◎
Longest run18 kmNew durable distance
EXAMPLE YEARLook back and see what changed.
Illustrative86runs completed
642 kmtotal distance
8new PBs
5level upgrades
10 wksbest pain-free streak
A year should tell a story: where you started, what improved, what interrupted training and how much stronger or more durable you became.
WHAT WORKS TODAY
Proof, without pretending the beta is finished.
RunAI already has working product foundations. Forecasting and some direct device integrations are still being developed and validated, so they are labelled separately from live capability and external research.
Connect an athlete account and bring completed running activity into RunAI for the training loop.
Planned sessions can be created, changed and replanned as the athlete's calendar evolves.
Completed runs can capture effort and niggle feedback rather than treating a completed activity as the whole story.
The forecast is an estimate of likely direction from the athlete and training ahead — not a guaranteed race result or a validated accuracy claim. It should update as evidence and the plan change.
External studies help shape training rules. They do not prove RunAI's own predictive accuracy or athlete outcomes.
We are not publishing synthetic testimonials, athlete counts or performance-gain claims. Product outcomes need real beta evidence first.
YOUR TRAINING DATAClear status, clear boundaries.
Data & privacy →RunAI uses connected training data and athlete-provided feedback to operate the product. Integration access can be disconnected from the athlete account.
RESEARCH BACKEDBuilt on evidence.
Built on evidence.
Kept up to date.
RunAI is grounded in sports science and exercise-physiology research. We continuously review new evidence, use it to challenge our assumptions and update the product when better evidence emerges.
Where a recommendation depends on research, the goal is to make the evidence understandable and link back to the underlying studies — not hide behind jargon or vague claims of “AI”.
REVIEWContinuously review new research
Training adaptation, endurance performance, injury risk, recovery and athlete monitoring.
TRANSLATETurn evidence into useful rules
Plain language first. No PMID soup dumped on the athlete.
TESTCompare research with real athlete outcomes
Research informs the system; actual athlete response determines whether it works.
UPDATEChange when the evidence changes
The model should get better rather than fossilise old assumptions.
01
HIIT for recreational endurance runners
02A systematic review found that combining high-intensity intervals with continuous running can improve performance, VO₂max and running economy.
Read the study ↗Specific sprint training improves sprint performance
03A review of 48 studies found specific sprint training produced meaningful improvements across sprint distances.
Read the study ↗Training load, fitness and injury risk
A review argues that appropriate progressive loading can build resilience, while rapid or poorly managed changes in load can increase risk.
Read the study ↗INTEGRATIONS
Your data should follow you.
Strava is already connected and working. Direct Garmin, COROS and Apple Watch integrations are in progress.
STRAVA · LIVEGARMIN · IN PROGRESSCOROS · IN PROGRESSAPPLE WATCH · IN PROGRESSINTERVALS.ICU · BRIDGE
Early-access bridge
Intervals.icu can also be used as a practical bridge for supported data during testing while the remaining direct integrations are built.
FAQ
Common questions.
How is RunAI different from a normal training plan?+
A normal plan mainly tells you what to do. RunAI is being built to understand the athlete, predict what planned training is likely to do, adapt when new evidence arrives, and make progress visible over time.
What does RPE mean?+
RPE means Rate of Perceived Exertion: simply how hard a workout felt to you. If a session that should feel manageable feels extremely hard, that is useful information the next plan decision should consider.
What does LT mean?+
LT means lactate threshold. In plain English, it is the boundary around a hard but sustainable running intensity. RunAI uses threshold evidence to calibrate training paces and to detect meaningful improvements over time.
How does RunAI adapt training?+
It combines completed training with athlete feedback such as pain or niggles, fatigue, RPE, Journal entries and missed or changed sessions. The aim is to reconsider what comes next instead of blindly following the original calendar.
Does RunAI work with Garmin, Strava and other platforms?+
Strava is already connected and working. Direct Garmin, COROS and Apple Watch integrations are still in progress. During early testing, Intervals.icu can also be used as a bridge for supported data where needed.
Does RunAI prevent injuries?+
No app can promise that. RunAI is being designed to respond to pain, load and athlete feedback so training can be adjusted when the evidence suggests backing off or redirecting, with the goal of reducing avoidable risk while still progressing.
How does RunAI use research?+
Research is not a badge on the website. We review relevant sports-science evidence, translate it into product rules, compare those rules with real athlete outcomes and update the system when better evidence emerges.
Is the forecast a guaranteed result?+
No. A forecast is an estimate based on the athlete and the training ahead. It should show uncertainty and change when the plan or athlete evidence changes.
JOIN THE EARLY BETA WAITLIST
A smarter plan should know you better every week.
Leave your name and email to hear about beta progress and opportunities to help shape RunAI. Pricing has not been announced yet.
After you submit, you'll only hear from us when there is a meaningful RunAI update or early-testing opportunity.