I’ve Used AI-Based Calorie Counting Apps and They’re Even Worse Than I Expected

The promise is tempting: Snap a photo of your food, and artificial intelligence will instantly tell you exactly how many calories you’re consuming. No more tedious manual recording, no more guesswork about portion sizes, no more human error. Apps like Cal AI, Lose It!, and MyFitnessPal’s new photo features claim to revolutionize calorie tracking by letting your smartphone camera do the heavy lifting.

But as someone with a long and complicated history of calorie counting (and, admittedly, a fair amount of experience at it) , I can tell you that counting calories from a photo is exactly as stupid as it sounds.

How AI Calorie Counting Should Work

Calorie-counting apps promise to solve what developers call the biggest problem in calorie tracking: human error. The pitch is compelling—why waste time searching through databases and measuring portions when your phone can instantly analyze your plate?

Apps like Cal AI or SnapCalorie AI use visual cues like color, texture, and relative size to make educated guesses about what you’re eating and how much.

They claim that AI techniques can solve the pesky problem of human accuracy in estimating calories, which is, frankly, easy to get wrong . Cal AI bills itself as one of the most sophisticated options in the field, so I decided to see for myself. The app was free for the first three days, then $29.99 per year.

The setup process is simple: Download the app, create an account, enter some basic demographic information, and set your goals. Here’s where I encountered my first red flag. The app cheerfully informed me that “losing 10 pounds is a realistic goal” — except that losing 10 pounds would actually move me into the underweight zone on my BMI. That kind of blanket statement shows a disturbing lack of nuance regarding individual health needs.

The Cal AI photo registration process consists of the following steps:

1. Take a clear photo of the food, preferably against a plain background.

2. Make sure all ingredients are visible and well lit.

3. For scale, use some reference object (such as a coin or your hand).

4. Upload your image and wait for AI analysis.

5. Please review and correct your app identification and portion estimates.

6. Keep a note in your daily journal.

The app gives detailed advice for achieving the best results: use natural light, avoid shadows, keep the camera parallel to the plate, and make sure ingredients aren’t hidden. These recommendations sound reasonable in theory, but they hint at a fundamental problem these apps face: the complexity of real-world nutrition.

Reality is wildly disappointing

I started my testing with something simple: a 222-gram Pink Lady apple. Of course, this would be an easy win for the AI—apples are among the most photographed foods on earth, with a distinctive shape and color that should be instantly recognizable.

Cal AI confidently identified my apple as tikka masala.

I gave it another try, this time taking a photo of an apple next to the barcode and placing it on a kitchen scale that showed its exact weight. This time, the app recognized it as an apple, but estimated it at 80 calories when the actual amount should have been closer to 120. That’s 33% less — not quite the accuracy you need if you’re trying to accurately track your intake.

The real test came with a more complex dish: my current lunch of fried tofu, onions, cucumbers, tomatoes, feta cheese, and chickpeas, all tossed generously in a homemade butter-based dressing. It’s the kind of mixed dish that seems to demonstrate the advantage of AI over manual data entry—no need to hunt down individual ingredients or estimate quantities.

The results were a masterclass in algorithmic overconfidence. The app identified the golden-brown fried tofu as croutons, which I had to manually correct. It was pretty good at identifying the vegetables and feta, but completely ignored the oil content. Even though the salad was clearly glistening with dressing, the app rated the entire dish at 450 calories.

This estimate was laughably low. One can of chickpeas contains about 400 calories , and my serving included about that, plus a significant amount of feta cheese and a few tablespoons of olive oil-based dressing. The actual calorie count for this dish would be closer to 800 to 900.

The app’s portion estimation proved even more problematic than identifying ingredients. When I photographed a smaller portion—less than a quarter of the original salad—Cal AI estimated it at 250 calories. According to the app’s own logic, less than 25% of the dish somehow contained more than 55% of its calories. The math just doesn’t work.

This highlights a fundamental limitation of photo-based calorie counting: Cameras capture two-dimensional images of three-dimensional objects. Without consistent reference points or sophisticated depth analysis, estimating volume from photographs is largely a guess. Even humans have difficulty with this task, which is why nutritionists typically recommend weighing foods for accuracy.

To get a more complete picture of the AI-powered calorie counting capabilities, I also tested two other popular apps: SnapCalorie and Calorie Mama.

SnapCalorie: Better Numbers, Same Problems

SnapCalorie immediately dispelled some of the skepticism by offering a much more reasonable daily calorie goal of 1,900 calories compared to Cal AI’s problematic weight loss messages. However, this accuracy comes at a high price — $79.99 per year after just a one-week free trial, making it the most expensive option I tested.

The app offers one interesting feature: an “add note” feature that lets you provide additional context about ingredients that the camera can’t see. In theory, this addresses one of the fundamental limitations of photo-based tracking.

When I tested SnapCalorie with the same Pink Lady apple, it did much better than Cal AI, estimating 115 calories. But the Greek salad test revealed familiar problems. SnapCalorie’s initial estimate was absurdly low, at 257 calories. When I photographed a smaller portion—the same quarter portion that had stumped Cal AI—SnapCalorie estimated it at 184 calories. The math still didn’t work; that smaller portion should have been about 25 percent of the larger portion, not 70 percent.

Deciding to give the app a fair chance, I used the notes feature to manually enter “a full container of tofu, feta, chickpeas, and olive oil.” Thanks to this human intervention, SnapCalorie bumped its estimate up to 761 calories — much more reasonable and accurate, though still not enough.

But this raises an obvious question: If I have to manually enter detailed ingredient information to get accurate results, what exactly is the photo doing? Essentially, I’m doing the work of traditional calorie counting while also going through the motions of taking a photo.

Calorie Mama: When AI Doesn’t Even Try

Calorie Mama provided the most frustrating and laughable experience of the three apps. The interface feels primitive, and the AI ​​performance is so poor that the app essentially abandons the premise of automatically analyzing photos.

Once you upload a photo, Calorie Mama requires you to manually confirm not only the food items, but also their portion sizes. This defeats the whole point of photo-based signups—you’re doing all the work that would have been required by manual entry anyway.

When I uploaded a photo of a Greek salad, Calorie Mama simply identified it as “tofu” — completely ignoring the vegetables, feta cheese, chickpeas, and dressing. The app then asked me to manually adjust the serving size and seemed to consider the registration complete, as if the complex mixed dish contained nothing more than simple tofu.

It wasn’t just inaccurate; it was useless. At least Cal AI and SnapCalorie tried to recognize a few ingredients, even if their calorie estimates were wrong. Calorie Mama seemed to abandon the core task entirely, reducing AI to little more than a useless photo storage system.

AI Calorie Counting Was Wasting My Time

The promise of AI-powered calorie counting is efficiency — click and go, no manual entry required. But my experience showed a different reality. I spent a lot of time correcting ingredient identification, adjusting portion sizes, and rethinking the app’s ratings. In many cases, I would have been faster using traditional manual entry via a food scale and database searches.

This creates an unfortunate dilemma: If you don’t carefully check the AI ​​results, you’ll end up with wildly inaccurate data. But if you check every record, you’ll lose the time-saving benefit that justified using the technology in the first place. It’s the worst of both worlds—the effort of manual tracking combined with the uncertainty of automated guesswork.

Perhaps most troubling is what happens when users don’t have the experience to recognize inaccurate estimates. My years of calorie-counting experience — as problematic as it was — gave me the knowledge to recognize when Cal AI’s numbers were wrong. But what about the users who trust the technology?

Systematically underestimating calories can be especially harmful to people trying to lose weight, as it can make them believe they are eating less than they actually are. Conversely, overestimating can cause unnecessary restriction or anxiety about food. Either way, inaccurate data undermines the whole purpose of tracking.

The main problem with AI-powered calorie-counting apps isn’t just technical — it’s philosophical. These tools presume and reinforce the idea that accurate calorie tracking is necessary and healthy. But research shows that obsessive calorie counting can do more harm than good for many people.

Intuitive eating , which focuses on internal hunger and fullness cues rather than external indicators, has been shown to be a more sustainable and psychologically healthy approach to eating. This framework emphasizes developing a healthy relationship with food based on how it makes you feel, rather than achieving specific numerical goals.

For most people, understanding the general principles of a balanced diet—eating plenty of vegetables, choosing whole grains over refined ones, including enough protein—will provide better long-term results than meticulously counting calories.

Summary

AI-powered calorie-counting apps promise to solve the problem of human error in diet tracking, but they introduce new forms of inaccuracy while preserving many of the same old problems. If your goal is simply to get a rough estimate of the number of calories in common foods, these apps can be useful. But for those looking for precision in tracking their intake, traditional methods combined with a food scale remain more reliable.

More importantly, I’d question whether counting calories accurately serves your health goals at all. For many people, developing a more intuitive relationship with food—based on satisfaction, energy levels, and overall well-being rather than numerical goals—leads to better physical and mental health. Perhaps the old-fashioned approach of listening to your body works better than any algorithm.

More…