Cal AI vs MyFitnessPal: same owner, and you can still pay twice

You have both icons on the phone. One of them has been the diary for years. The other got installed because pointing a camera at a plate felt easier than searching a database at 9pm. The reasonable assumption, now the acquisition is public, is that they became one app. They did not.

MyFitnessPal's own announcement on 2 March 2026 says Cal AI keeps operating as a standalone product. TechCrunch reported the same week that the chief executive saw different markets, with no plan to fold the camera into the main app or move users across. The database did get shared. The subscription did not. The truth is slightly worse than a merger: you can pay both companies, which are now one company, and still be logging in neither by week ten.

Our bias, stated first

We make Pocket Fit. This page exists partly to sell it, so read every sentence about us as a claim from an interested party.

What we can check is the public record. The acquisition dates, the "still standalone" line, and the database size come from MyFitnessPal's 2 March 2026 announcement and from TechCrunch the same day. Cal AI's own FAQ says the estimator is about 80% accurate. MyFitnessPal's help centre has said since 1 October 2022 that barcode scan sits behind Premium. Prices move by store and by promotion. Where a figure is not on a page we read, we do not invent it.

One honesty tax up front. Cal AI is better than a database at getting a reluctant person to log a mixed plate. That is a real product achievement. Owning the database does not erase it, and our training app does not erase it either.

The short answer

These two apps now share an owner and a food catalogue. They still answer different questions.

Cal AI answers "what did I just eat" with a photo, a barcode, or a short description. MyFitnessPal answers "what is in this branded food" with a very large search box, and it charges Premium before the fast paths (barcode, Meal Scan, voice) will do the typing for you.

Choose Cal AI if the only log you will actually keep is a photo of a home-cooked or restaurant plate, and you can live with an estimate.

Choose MyFitnessPal if your week is packaged food, you want the long catalogue and the web diary, and you will pay Premium for the scanner.

Choose Pocket Fit if the reason you track food is the training week: protein on a cut, fuel for a heavy session, one place instead of two subscriptions. For the wider field, start with best food tracking app. For the head to head against the camera alone, read Pocket Fit vs Cal AI.

Cal AI vs MyFitnessPal side by side

Figures below are from the company posts, the FAQ, the help centre, and store listings as read for this piece in 2026. Your store can differ.

The row that should change your budget is the last one. The acquisition merged the companies. It did not merge the bills.

What the photo still gets wrong

This is the part that did not change when the press release went out.

Shonkoff and colleagues reviewed 52 papers for the Annals of Medicine, comparing fully automated image-based dietary assessment with ground truth. They could not pool a single accuracy number. Relative calorie error, where they could extract it, ran from 0.10% to 38.3%. Error was smaller for one plain food and larger when several foods sat in the same frame. Volume error sat in a similar spread, 0.09% to 33%.

Cal AI's "about 80% accurate" lives inside that range. It is a fair self-description, not a boast, and it was true before MyFitnessPal bought the app. A bigger database fixes the long tail of branded foods. It does not photograph the oil you cannot see.

Why the diary still dies in week ten

A perfect estimate you abandon is worth less than a rough one you keep. That is the finding that should decide this comparison, and it is not about which logo is on the camera.

Burke, Wang and Sevick reviewed 22 studies for the Journal of the American Dietetic Association and found a consistent association between dietary self-monitoring and weight loss, while calling the methods weak. Turner-McGrievy and colleagues then looked at two six-month mobile trials, 124 people in total, in the Journal of the Academy of Nutrition and Dietetics. Fewer than half were still tracking after week 10. The days on which someone logged at least two eating occasions explained the most variance in weight loss.

Read that against two subscriptions. If barcode scan is how you log a supermarket week, and MyFitnessPal keeps it behind Premium, the free diary is the slow diary. If Cal AI's camera is the only gesture you will repeat, paying for MyFitnessPal's search box on top of it does not buy adherence. It buys a second icon.

Where the evidence runs thin

We do not have a trial of Cal AI against MyFitnessPal in the same people. Shonkoff's review could not meta-analyse the photo papers, because the image sets and the ground truths did not match. The 0.10% to 38.3% range is a range, not a score for either app. Cal AI's 80% line is the company's own FAQ, not an independent lab.

Burke's review is old, and the authors said the self-monitoring evidence was methodologically weak. Turner-McGrievy's 124 people are two mobile trials, not a photograph of your kitchen. None of these papers tested a post-acquisition product. The deal terms were not disclosed, so anyone quoting a purchase price is guessing, and we will not.

Our third-option pitch also flatters us. Pocket Fit does not have MyFitnessPal's catalogue, and we do not claim Cal AI's one-gesture onboarding. If you never lift, a training membership is the wrong answer to a food question.

What to do with the two apps you already downloaded

• If you only log restaurant plates, keep Cal AI and stop paying for a database you do not search.
• If you only log barcodes, pay for MyFitnessPal Premium once, and do not add a photo subscription on top.
• If you pay for both, pick the one you opened on three days last week and cancel the other. Week 10 is the test, not the feature table.
• If you lift, put food in the same place as the programme. A calorie number with nowhere to go does not change Thursday's session.

Fuel in Pocket Fit is that place: protein, carbohydrate and fat, photo or barcode, no red numbers and no shame. The AI coach can rebuild a session from a sentence. The scheduler moves a missed day into the rest of the week instead of deleting it. Body budget keeps the workout, the streak, sleep and nutrition in one tally. The programme updates week to week from what you log.

For how wrong a camera can be on a mixed plate, read how accurate photo calorie apps are. If the real alternative you are weighing is an adaptive macro app rather than a camera, read MacroFactor vs MyFitnessPal.

Put it together

Same owner. Two jobs. Two bills. The photo is still an estimate, somewhere between almost exact and more than a third out, and the diary still only works while you keep it past week ten.

Pocket Fit is the option when training is the reason you opened a food app at all. Georgi, who built it, went from 122 kg to winning first place at The Yard Games, having lost 38 kg along the way. The app exists because of the bad weeks, not the good ones. That story is at our story.

Cal AI vs MyFitnessPal: common questions

Did MyFitnessPal buy Cal AI?

Yes. The deal closed in December 2025 and was announced on 2 March 2026. TechCrunch reported a retained team of seven, terms not disclosed, and Cal AI already connected to MyFitnessPal's database of 20 million foods, 68,500 brands and 380-plus restaurant chains.

Are Cal AI and MyFitnessPal the same app now?

No. MyFitnessPal said Cal AI continues as a standalone product, and that it had no plan to merge the apps or migrate users. A MyFitnessPal Premium subscription does not unlock Cal AI, and the other way around is the same.

How accurate is a photo calorie estimate?

Shonkoff and colleagues could not produce one number from 52 papers. Relative calorie error ran from 0.10% to 38.3%, smaller on a single plain food and larger on a mixed plate. Cal AI's own FAQ says about 80% accurate, and that no food app is perfect.

Is MyFitnessPal's barcode scanner free?

Not since 1 October 2022. The help centre puts barcode scan, Meal Scan and voice logging on Premium. Manual search of the catalogue is the free path. Premium is commonly around $19.99 a month or $79.99 a year, and your store can differ.

Which one should you pay for if you also lift?

Neither, if the point of the log is the training week. Fewer than half of the people in Turner-McGrievy's mobile trials were still tracking after week 10. Pocket Fit keeps Fuel beside the programme, the scheduler and Body budget, so the number has a session to change.