Pocket Fit — AI workout+food app whose hook is 'see your future…
Pocket Fit (PH launch, ~22 votes, two-person London team): one photo → AI projections at 1/2/3/6 months plus a 3D avatar that evolves as you train, on top of a full workout+food app — filed as evidence that the future-body-visualization family is piling up, not as an open gap.
What is the idea?
Full-stack fitness app (auto-progressed strength plans, photo food logging, muscle coverage, mood, streaks, social roasting) whose differentiating hook is AI future-body projection plus an avatar that levels with real training.
Why would anyone bother?
— (filed as family evidence, not a build candidate as-is)
What is missing today?
The gap its copy implies — 'nobody makes progress visible' — did not survive one counter-search: FutureSelf, FitTransform, Body Goal AI, UpliftPal, seemyfuturebody and a pile of web fat-to-fit filters all shipped the same projection mechanic within roughly the last half year. FutureSelf even claims the loop (real progress photos tracked against the AI image).
What makes this one different?
The only cut still unclaimed as far as we can see: own the reconciliation, not the fantasy — weekly same-pose re-shoots aligned and replayed against a *reachable* projection, sold on honesty ('what actually changed') rather than hope. CN side unchecked: 小红书's manual 对比照 culture is the native behavior; whether anyone owns it in Chinese is the one open question.
What is the sharpest thing about it?
It bundles the projection into a real training loop (avatar fed by actual workouts) where most of the family are one-shot filters.
What could kill it?
Projection is hope-selling: the month the mirror diverges from the AI image, motivation doesn't dip — it snaps, and the diet-industry pattern says it snaps hard. The pile-up has the same shape as the prompt-tool wave: an image-gen capability window opened and everyone jumped. And the mechanic is one feature away for every incumbent (Keep, Zing, Nike).
Who pays, and for what?
If the honest-reconciliation cut works: people in weeks 3–12 of training, paying for proof over promises.
What is the main risk?
Body-image backfire is intrinsic to the whole family — an AI image you never reach is worse than a number.
How much work is it?
about two months of work
Where does the evidence come from?
https://www.pocket-fit.app/
← Soloop — approval-gated 'agent OS' for solo founders (PH #2) · Kindergarten classroom display work →