Every tracker guesses.
We’re the one that admits it.
Photo calorie apps love a confident number. The science says that confidence is theater — so we built the opposite. This page is the long version, for people who want to know exactly what they’re trusting.
What the research actually says
25–35%
That’s the average calorie error independent, peer-reviewed studies measured in the best photo-AI systems — and they tend to underestimate, especially on big portions. Anyone telling you otherwise is selling something.
Portion size is the hardest part: even professional dietitians can’t judge it reliably from a photo alone. Neither can an AI. That’s why BuonApp makes it effortless for you to supply the one thing the photo can’t show — one line of text, one slider.
Sources: three peer-reviewed studies, linked here before launch [1] [2] [3].
How honesty looks in the app
- The range is always there. “584 kcal — likely 513–654.” Every estimate carries its honest range, right under the number.
- Confidence, out loud. High, medium, or “low — please review.” When it’s guessing, it says so instead of pretending.
- Assumptions you can correct. “Olive oil pooled on the plate suggests generous use.” You fix the guess itself — not just the total.
It even tells you about the lemon
“White cloth parcel contains a lemon half — not counted as a caloric item.”
From a real analysis. That’s the level of honesty we think a food diary owes you: every estimate ships its full reasoning — what it saw, what it assumed, what it deliberately didn’t count.
It’s honest about your burn, too
Your plan assumes a certain daily burn from the questionnaire — but real weeks have skipped workouts and sick days. BuonApp records what your devices actually measured (Whoop, Apple Health, or your phone’s steps) and reconciles: the day review shows the measured energy balance next to the plan’s assumption, and the week review says it plainly — “your plan assumed ≈16,500 kcal out; your devices measured ≈14,800 — the real deficit was smaller than planned.” Plans flatter; measurements don’t — we show you both.
Where it struggles — and what happens then
A number, or nothing
We haven’t finished our own ground-truth accuracy study yet. Until it’s done, we won’t quote you an accuracy number — ours or anyone’s marketing department’s. Here’s exactly what we’re running:
- 50+ meals, weighed and measured before being photographed
- Diverse dishes — not just the easy, photogenic ones
- The pass/fail bar set before we see the results
- Results published here — pass or fail
Full honesty, since that’s the brand: today’s ranges come from the AI’s own confidence in what it sees — an honest estimate, not a laboratory-calibrated interval. The study above is how we hold it to account. And as of our June 2026 review of the major trackers, none of them showed you any of this at all.