Know if you're getting stronger. Know what to do next.
Most apps record your training. Repcurve tells you whether you are getting stronger.
Log every set. Repcurve reads your own training history and tells you where each lift stands — progressing, holding or declining — what is behind it, and what to change next. No spreadsheets, and no guessing: the number sits under every claim.
The trace
Squat e1RM · 90 days148 kg ▲ +0.9 kg/wk
n=14 · 90 days · 90% CI ±0.3 kg/wk · ▒ gap marked
One lift, ninety days, read the way an instrument reads it.
Every working set becomes an estimated one-rep max. A line is fitted through them with a confidence interval, and the slope of that line is the verdict. Gaps are marked, never smoothed over. A record is a hairline and a label, not a star.
The verdict
Every lift gets a clear verdict.
Repcurve reads your recent training and says what is happening to each lift. Progressing means strength is going up. Holding means it has stalled. Declining means it is going backwards.
Every verdict is backed by your own training data, and the chart that produced it sits directly beneath the sentence.
- Progressing
-
Load is moving up at stable effort.
e1RM +0.9 kg/wk · RPE stable
- Holding
-
Flat inside the interval. Not yet a plateau.
slope within CI of zero · RPE stable
- Declining
-
Trending down, or holding at rising effort.
e1RM −0.6 kg/wk · RPE rising
One session
Log it. Read it. Ask it why.
01 · Log
Weight, reps, RPE. Nothing else.
Three figures per set, tapped in between sets. Warm-ups are marked and kept out of every aggregate. The live e1RM strip updates as you go, and the whole screen works with no signal at all — a set is written to the device the moment it is logged.
02 · Read
The verdict, then the evidence.
Every session ends with one sentence a coach could say out loud, the four numbers behind it, the chart that proves it, and at most two flags worth a look. The number sits under every claim.
03 · Trend
Ninety days, per lift.
Each exercise gets its own charts: the e1RM trend with its fit and interval, the load you move at a fixed effort, and rep quality — the same load, this month against last. A gap in training is marked, never smoothed over.
04 · Ask
Ask it why.
Chat is bound to the session. Answers quote the figures they were handed and cite them beneath the sentence. Where your data does not cover the question, it says so rather than guessing.
What to do next
A verdict is only useful if it changes something.
Knowing a lift has stalled is half an answer. Each session ends with one recommendation for the next time you train it — what the pattern in your data usually calls for, and why.
It will not write your programme, and it will never programme around pain. It tells you what your data shows and what typically follows from it. The decision stays yours.
- Squat Progressing
- Working. Leave the load progression as it is.
- Bench press Holding
- Effort is stable, so the stimulus has gone stale. Add a set, or change the variation.
- Deadlift Declining
- Effort is rising as load falls. Consistent with fatigue — deload and read it again next week.
What is measured
Five figures per set become a reading per lift.
Everything Repcurve derives is a function of exercise, weight, reps, RPE and a timestamp. Each reading has a minimum amount of history before it is asserted, and both gates — sessions and weeks — are shown until they are met.
- Plateau
- Fatigue-masked progress
- True regression
- Silent exercise dropout
- Intensity monotony
- Load spike
- Readiness divergence
Honest limits
When there isn't enough data, it says so.
No invented insights. Until there is enough history to be sure, Repcurve shows the sessions it has, counts the ones it still needs, and withholds the verdict. A guess is never rounded up into a finding.
The same rule governs the language. Claims are hedged in proportion to the interval — "consistent with", "the likeliest explanation is" — never "this is because".
99 kg
What it will not do
It answers from your data, and stops where the data stops.
Why has my bench stalled? Am I recovering? Ask at the end of a session and Repcurve answers from your own figures, quoting the ones it was given rather than inventing any. Where your data does not cover the question, it says so.
It will not programme around pain — it will tell you to reduce load and see a physiotherapist. It does not set calorie targets or goal weights, and it treats medical questions as out of scope.
Why has my bench stalled?
Bench e1RM has been flat for three sessions while average RPE rose from 7.5 to 8.5. That pattern is usually effort creeping up under a load that is not moving, rather than a strength loss.
e1RM +0.1kg/wk · 90% CI −0.4 to +0.6 · n=9 · 90 days
Principles
Built to four rules that do not have settings.
Offline, always
A set is written to the device the moment you log it and syncs later. Nothing in the logging path waits on a network.
One baseline: yours
No leaderboards, percentiles or benchmarks — not as a setting, they do not exist. Every comparison is against your own history.
Nothing asserted early
Below the data threshold it shows what it has and counts what it still needs. A guess is never rounded up into a finding.
Regression is information
Declining is amber, not red — a reading to act on, not a failure. No alarm, no apology, no exclamation marks.
Pricing
Four weeks free. Then you are paying to understand your training, not to log it.
Free trial
Four weeks
The whole product, not a sample of it — every chart, every finding, the chat. Logging works fully offline; nothing in that path waits on a network. No card is needed to start.
Then
£9.99 per month
One price, monthly, cancel any time. If you stop, the app closes behind the paywall and your logged data stays exportable — it is never held hostage.
When the trial lapses the app closes behind a paywall, and the paywall keeps its manners: the finding is given away in full, only the diagnosis is withheld, the price is stated plainly and "not now" sits at equal weight.
Questions
Know if you're getting stronger. Know what to do next.