Cal AI made AI calorie tracking mainstream and most apps since then just copy the scan-and-guess formula and call it a day.
Our team built Dietfit to fix what that formula misses. In a 30-plate head-to-head test, our scan beat Cal AI at 80.1% accuracy - but the bigger difference is what happens after the scan. Every ingredient Dietfit detects shows up as an editable pop-up right on your photo. Miscounted the dressing? You just tap it and fix it right away.
Streaks, widgets, Apple Health sync - we have all of them in 1. But the core bet is this: a calorie tracker should get smarter about you over time, not just scan and forget.
I have a complicated history with tracking tbh. This one feels less obsessive because it's fast and doesn't shame you, and it matters more than people think. And congrats on the launch!
@do_mthuy this honestly made our day, thank you. you’re right that it matters more than people think. we want Dietfit to feel like a helpful nudge, never a judgment, and we’ll keep that in mind as we build. 🙏
Hi Product Hunt!
One thing we've learned while building Dietfit: a calorie estimate shouldn't look more certain than it is.
A number like “537 calories” looks precise. But a food photo doesn't tell you everything: how much oil went into the pan, what's inside a sauce, or exactly how large the portion is.
A photo scan is an estimate, not an exact measurement. That's why we let users review the ingredients Dietfit detects and edit them directly on the photo.
It doesn't remove every uncertainty, but it gives users more control instead of blindly accepting a number.
For us, it's not just about how quickly we can show a number, but how we help people understand it.
What would help you interpret an AI food estimate: a range, an explanation of the assumptions, or a clearer way to review the ingredients?
Would love to hear your thoughts!
Do you plan to do it for Android?
Clean, fast, and it actually seems accurate. Upvoted 🔥
@huy_che Cheers!!
Congrats on the launch! Macro breakdown straight from a photo is kinda wild. Didn't expect it to be this fast
Adding 1 data point to this: we actually ran a 30-plate head-to-head test against Cal AI before launch, Dietfit came out at 82% accuracy. The scan-and-edit flow Ella mentioned is what made the biggest difference for us during testing, way faster than hunting through a food database to fix one wrong item.
Excited to see this out in the world today 🧡
Most people don’t stop tracking calories because they stop caring about what they eat. They stop because keeping track of everything becomes another task they have to manage.
Typing every ingredient, estimating portions, and repeating the same process meal after meal can get exhausting.
That’s one of the problems we wanted Dietfit to address. You can scan a meal, review the ingredients Dietfit detects, and edit individual items directly on the photo instead of having to accept one final number.
We’re also trying to make the experience useful beyond logging, with Sara for personalized guidance, plus widgets, streaks, and Apple Health sync.
Our goal isn’t to make people track more. It’s to make tracking easier to keep up with.
What’s the biggest reason you’ve stopped using a calorie tracker in the past?