Abstract
BitePal is an AI-photo calorie tracker with a gamified mascot, but a recent update forced all food logging through an unreliable "Analyze with AI" step, removing the fast manual entry long-time users relied on.
Background
BitePal: AI Calorie Counter tracks calories and macros from a food photo, paired with a gamified raccoon mascot.
What it does
An AI-photo calorie/macro tracker with a gamified mascot, fasting timer, and wearable sync.
What changed
A recent update forced all food logging through an "Analyze with AI" step and removed the quick manual-entry flow users relied on for accuracy.
Why it is soft
Multiple long-time users say the app "used to be" accurate and fast before this specific update — the gap is a regression, not a fundamentally broken idea.
Strengths
The core loop — photo, estimate, gamified feedback — clearly works for many: several reviewers report real, sustained weight loss using it, and the mascot/streak mechanic is repeatedly cited as what keeps them logging daily.
Market gap
The recurring complaints, most-common first:
Manual entry removed
A recent update removed or hid fast manual entry, forcing a slower and less accurate AI-analyze flow — the most common recent complaint from previously happy long-term users.
AI recognition inaccuracy
Portion sizes and non-American foods frequently misread, requiring manual correction most of the time anyway.
Cancellation dark patterns
No in-app cancel button, unresponsive support, and charges continuing after cancellation attempts.
Unreliable fitness sync
Step counts and Health Connect data reported wildly undercounted or duplicated.
New update ruined this app. Before I could have just typed in what I had but the new "analyze with AI" feature makes it so unnecessarily complicated and it's no where near as accurate.
There is no way to cancel subscription in app. been charged another £57 despite terminating my account.
I have a 142 day streak going, but the recent ai updates have been terrible. Manual entry has been changed to require an ai analysis, and if the ai yields no results, you can't log it.
- Keep one-tap manual entry alongside the camera. The update that removed it is the single most-cited reason long-term users are leaving.
- Let a corrected estimate be remembered per-food. So users aren't re-guessing the same dish every time.
- Add a visible in-app cancel-subscription button. The "no way to cancel" complaint pattern is costing trust independently of the AI issues.
Build complexity
Estimated for one developer, part-time, no funding, no paid acquisition.
| Workstream | Weeks |
|---|---|
| Manual + camera food logging with editable entries | 2.0 |
| AI photo recognition (off-the-shelf vision API) | 1.5 |
| Fasting timer + water/macro tracking | 1.0 |
| Subscription billing with self-service cancellation | 1.5 |
| Total | 6.0 |
Expected revenue
Assumes an $8.99/month subscription in a crowded calorie-tracking category, so conversion is modest even with a fixed regression.
| Month 12, monthly | ||
|---|---|---|
| Installs | 6,500 | |
| Activated | 3,200 | 49% |
| Still on at 30d | 1,700 | 26% |
| Paid | 300 | 5% |
| Revenue | $2,700 |
Biggest challenges
The calorie-tracking category is extremely crowded with well-funded incumbents (MyFitnessPal, Cronometer, Lose It, and near-identical AI-photo trackers) — a solo builder's realistic wedge is narrow: users specifically burned by this one update, not the broader market, so growth depends on catching switcher search traffic rather than broad acquisition.
AI angle
An off-the-shelf vision-language model API can do photo-based calorie estimation for a few cents a call — this is what BitePal itself already does — so the AI angle is table stakes, not a genuine edge; the actual opportunity is UX (keeping manual entry) and billing honesty.
Conclusion
Build it. A specific, verifiable regression to fix, though the crowded category and thin AI differentiation keep the margin modest.
6 weeks
Time to v1
$8.99/mo, manual entry always available
Price and model
$2,700
Revenue at month 12
A 142-day streak user calling the AI update worse than before
The fact that decided it