There’s a fight in the lifting world right now, and it’s easy to take the wrong side of it. flexRep takes a specific one: pro-evidence, anti-cargo-cult, and quietly honest about what any number can actually tell you. We cite the literature and state its limits — and we hand you the read, because the call is yours.
We'll let you know when there's room.
For a decade, “science-based” was the prestige position in fitness: cite a study, win the argument. By 2026 that’s curdled. People are tired of the constant flip-flopping, the screenshot-an-abstract authority, the thirty-minute setup to justify one exercise, the sense that every session now comes with homework. The criticism is largely fair — a lot of “science-based” content is cargo-culting: a single study, often unread past the abstract, waved around to prove a point the creator already held.
But the backlash contains a trap. The valid critique is of the cargo-culting — the bad citation, the false precision, the manufactured authority. It is not a critique of evidence itself. Those two things are being conflated fast, and the conflation is part of a larger post-truth reflex: because some experts overreached, expertise is suspect; because one study was wrong, studies are worthless. That reflex feels like wisdom, and it throws out the useful thing along with the annoying one.
flexRep refuses both poles. We won’t defend the cargo-cult, and we won’t join the bonfire of evidence to win the crowd that’s sick of it. We hold the harder, less viral line: cite the literature, and state its limits. Both halves. Every time.
Not a vibe. These govern the copy, the charts, and the code — and they’re written down in the product’s design documents so a future version of us can’t quietly walk them back.
Every formula traces to a published source, and every methodology card names what that source does not support: the population it was validated on, the rep range it breaks down outside of, the disagreement it papers over. A citation without its limits is the exact move the backlash is rightly tired of. We ship both halves, every time.
An eight-week study on twenty undergraduates is weaker evidence about your training than your own six-month e1RM trend. The charts describe your record. The AI paraphrases what the diagnostics found in it. Neither tells you what “the science says” you must do. You remain the author of your training.
The on-device AI reads your week back to you — and it is mechanically prevented from inventing a number. It offers, at most, one optional suggestion tied to a signal you can verify. It never issues a verdict from authority. “Coach” is a word we reserve for the real human you share your data with.
No “you’re leaving gains on the table.” No streak guilt, no nagging, no implied failure for not chasing every signal. Depth is opt-in — the Simple view is just Log, History, Body until you reach for more. We respect evidence; we will not weaponize it against your peace of mind.
The fastest way to betray everything above is to let the app — or its marketing — start selling certainty. So these don’t ship, in any channel.
A philosophy is cheap if the product doesn’t enforce it. Here’s where it’s wired into the surface — each one provable on its own page.
Formula, source, year, and named limitations — one tap from the chart itself.
See the methodology →A regex validator checks every figure against the diagnostic payload before the card ships. No invented stats.
How the pipeline works →“Not enough data yet” is a real answer. The surface names what’s missing instead of guessing.
The charts, in detail →No ads, no nags, no manufactured anxiety — the commitments we publish so you can hold us to them.
What we refuse to be →flexRep keeps that record honest, and hands it back. You make the call.
We'll let you know when there's room.