Abstract

MenuFit built a genuinely wanted product — tell me what to order at any restaurant for my diet — then buried it behind a survey-then-paywall funnel so aggressive that most of its recent long reviews never get past onboarding to judge the actual feature. Layered on top: nutrition numbers that don't match the restaurants' own published data, a subscription/restore flow that double-charges and locks people out of what they paid for, and no metric-unit or non-US-chain support despite a "worldwide" claim. None of these are hard problems — they're exactly the kind a solo builder can fix by not making the same choices.

Background

MenuFit (MenuFit - Healthy Eating Out) asks users their dietary goals and restrictions, then recommends what to order — or avoid — at a given restaurant, with a paid tier for full macro breakdowns and personalized meal plans.

Strengths

The underlying idea has real pull — reviewers repeatedly acknowledge the concept even while frustrated with the execution, and the app already has a UI reviewers call clean and an install base spanning people managing diabetes, GLP-1 medication diets, and ordinary fitness goals — demand a new entrant doesn't have to create, only capture.

Market gap

Complaints cluster into four groups, most-common first.

Payment before value. The largest and most repeated complaint by far: the app collects a full dietary survey and an email — and in many reported cases a card number — before showing a single real recommendation, with several reviewers calling this deceptive given the store listing shows the app as free.

Data accuracy and coverage. Calorie and macro figures reviewers cross-checked against a restaurant's own site frequently don't match; recommendations reportedly ignore stated allergies (gluten) and diet type (vegetarian); some major chains (a large coffee chain) are missing entirely despite the "every restaurant in the world" claim.

Technical reliability. Google sign-in loops, an AI chat feature that reviewers say crashes the app, and — most damaging commercially — a subscription/restore flow that repeatedly locks paying users out or charges them twice.

No international support. Imperial units only, US-centric chain menus, and no way to switch to metric — reviewers outside the US describe the "worldwide" claim as effectively false for them.

1. Let a user see several real, personalized recommendations before any account or payment step — the single change most likely to defuse the largest complaint cluster.
2. Source nutrition numbers directly from each chain's own published nutrition data and mark anything not verified, rather than presenting a guess as fact.
3. Ship with metric units and treat non-US coverage as a real launch requirement, not an afterthought behind an unqualified "worldwide" claim.

Build complexity

Assumes one part-time solo developer, no funding, no team — the estimate below is deliberately conservative because the nutrition dataset, not the app itself, is the actual bottleneck.

Expected revenue

Assumes pricing similar to the incumbent's already-validated $9.99/month (or a discounted annual plan), but with a funnel that shows value before payment — a meaningfully lower activation/paid-conversion rate than a forced-paywall app reports on paper, but one that isn't undermined by refund disputes and 1-star pileups.

Biggest challenges

The hard part isn't the app, it's the data: there is no official public API for restaurant nutrition information, so building and then keeping current a menu/macro database for even a few dozen chains means either manual upkeep or a scraping/parsing pipeline that has to be re-validated every time a chain changes its menu. A solo builder also inherits a trust deficit the whole sub-category has earned — reviewers already use the word "scam" for this exact kind of app — so a new entrant has to visibly behave differently (free look before payment, sourced data) to even get evaluated on its merits.

AI angle

There's a genuine angle here, and it's specifically in the bottleneck identified above: turning a chain's published nutrition PDF or menu page into structured, queryable data is exactly the kind of extraction work an LLM does far faster than manual entry, which directly shrinks the dataset workstream in build_complexity. The incumbent's own AI chat feature is what reviewers say crashes and "makes up menu items" — the fix isn't avoiding AI, it's grounding it strictly in the sourced dataset rather than letting it free-generate, turning the incumbent's weakest feature into a differentiator instead of dropping it.

Conclusion

Build it. MenuFit proves the demand and hands a competitor a documented list of exactly what to fix: show value before payment, source real nutrition data, and don't break the subscription flow. None of that requires new technology, just not repeating the incumbent's choices.