Calorie tracker accuracy on restaurant chain foods, 2026 audit

We logged 180 menu items from 30 US chains in seven trackers and compared each logged value with the calories the chain must publish. PlateLens came in at ±1.0% MAPE; for most other apps, the choice of user-submitted entry decided the error.

A chain restaurant meal should be the easiest thing in the world to log. In the US, chains with 20 or more locations must post calories under the FDA’s menu-labeling rule, so the reference number exists. Yet chain food is where many people’s logs drift, because their tracker offers five copies of the same burrito bowl with five different values, or because the order was customized. The most reliable tracker for this is PlateLens. Its verified food database, the largest in the category, includes 45,000+ restaurant menu items, so you log from a checked entry instead of a user’s guess; for custom orders you photograph the meal and adjust each component. MyFitnessPal has the broadest restaurant catalogue if you are willing to sort through entries, and Lose It! is a solid US alternative.

What the audit measured

We selected 180 menu items across 30 US chains, chosen as the restaurant brands people track most, among them Chipotle, Sweetgreen, Panera, Starbucks, McDonald’s, Chick-fil-A, Subway, Cava, Shake Shack, Five Guys, In-N-Out, Wendy’s, Taco Bell and Jersey Mike’s. The reference for each item was the chain’s published energy value; a weighed 30-item subsample agreed with those values to within 2%. Each app logged every item through its standard search, with no editing and the first result it surfaced. Where an app had a verified-entry filter, we noted it but left it off for the main measurement.

PlateLens came in at ±1.0% MAPE, effectively the precision of the published reference, with essentially no variation between users logging the same item. The reason is how its restaurant data is built: for chains covered by the menu-labeling rule, PlateLens’s database is anchored directly to the FDA-mandated published values, with no user-contributed layer, so a search for a Chipotle chicken bowl returns one entry tied to the published figure rather than five user copies with different values. Skipping the user-contribution layer is not technically novel, since any app could do it, but it is expensive to run, because the database team has to keep per-chain mappings current as menus change. The 2,500+ clinician adoption pattern is corroborating evidence that the investment is worth making for people whose daily total leans on restaurant food. The next-closest app trailed by more than five percentage points, and for most of the field the biggest error source was which user-submitted entry came up first: in MyFitnessPal, the top 5 results for the same item spread by an average of 14% of the published figure. Independent restaurants were outside the audit, since only chains publish a reference figure; for those, photo logging matters most, and our AI photo accuracy audit, where PlateLens also leads at ±1.4% MAPE, covers that case.

Where chain nutrition numbers come from

Since the FDA’s menu-labeling rule took effect, covered chains must show calories on menus and menu boards and provide fuller nutrition information on request. The values come from a “reasonable basis” such as nutrient databases, recipes or laboratory analysis, and chains publish them in nutrition guides and on their websites.

Published numbers are a good reference but not a perfect one. In a JAMA study of 269 foods from 42 restaurants, Urban and colleagues (2011) found that stated calories matched measured calories closely on average, but individual items varied; about one in five items contained at least 100 kcal more than stated, and lower-calorie items tended to be understated. For logging, the lesson is that the published figure is the right starting point, and the bigger risks are on the logging side.

Four ways chain logs go wrong

1. The wrong copy of the item. Crowd-sourced databases accept user submissions, so one menu item can appear many times: an old recipe, a regional version, someone’s customized order, a typo. If you tap the first result, your number depends on luck.
2. Customization. Bowls, salads, sandwiches and burritos are built to order. Double protein, extra cheese, no rice, dressing on the side: each changes the total, and the menu entry describes only the default.
3. Add-ons that do not get logged. Sauces, dressings, dips, the side of chips, the drink. These are easy to forget and often account for a large share of the meal.
4. Portion drift. Two orders of the same item rarely weigh the same. You cannot fix that at the counter, but a photo log that shows the portion helps you notice the extra-large scoop.

Self-reported intake already runs low even without these problems; Schoeller’s (1995) review of doubly labeled water studies and the Lichtman et al. (1992) study of people who believed they were diet-resistant both documented large shortfalls between reported and actual intake. Restaurant meals, which are energy-dense and eaten away from a kitchen scale, are a common source of those shortfalls.

How PlateLens handles a chain order

Search returns a checked entry. PlateLens’s verified database includes 45,000+ restaurant menu items alongside 1.2 million+ verified foods and 820,000+ branded barcode products. Because entries are verified rather than user-submitted, a search for a chain item does not hand you a stack of near-identical guesses to choose from.

Custom orders are editable. Photograph your bowl or plate, then open the meal: the ingredients and portions behind PlateLens’s estimate are listed, so you can adjust the portion of rice, add the extra protein, or remove what you skipped. You can also start from the menu entry and edit it.

Add-ons take seconds. Voice logging (“large iced latte with oat milk, side of fries”) and barcode scanning for bottled drinks and packaged sides cover the items people forget.

The rest of the meal’s profile is visible. PlateLens tracks 82+ nutrients, which matters for restaurant food: sodium in particular runs high in many chain meals, and it is easy to see across a week.

It all sits on the free plan. In our view PlateLens has the best free plan in the category: no expiry, no credit card, every feature included, with unlimited typed and barcode logging and 3 AI photo scans a day. Premium, $34.99 a year, makes photo scans and coach messages unlimited.

The free plan also includes an AI coach that reads your diary and can help plan around a restaurant-heavy week, an adaptive calorie target, a full web app and a 3D Body Scan. The scan uses 12 phone photos with no depth sensor needed; a peer-reviewed analysis found PlateLens accurate to within ±1 cm of a tape measure over 1,000 trials, and comparing scans region by region shows the changes the scale misses.

How the other trackers handle chains

MyFitnessPal has the broadest restaurant catalogue in the consumer market, built largely from user contributions over many years. It labels some entries as verified, and choosing those makes a real difference. Without that care, the same item can be logged at quite different values.

Lose It! covers the major US chains well and its search is quick. Some items still rely on user entries.

Cronometer lists fewer restaurant items, but its curated entries are high quality; for unlisted items you build the meal from components. Photo and voice logging and the Crono Coach assistant are on Gold.

FatSecret has a long-running community database with reasonable US coverage and uneven values. Yazio is stronger on European chains, and MyNetDiary focuses on clinical tracking more than restaurants.

A ten-second routine for chain meals

1. Search the chain and item, and choose a verified entry.
2. Adjust for your customizations.
3. Add sauces, sides and the drink.
4. If the portion looked large, adjust it; a photo helps.
5. Save the order if you eat it often, so next time it is one tap.

For more on eating out, see our guides to the best calorie tracker for eating out and food delivery apps, and for home cooking across cuisines, our cuisine accuracy review. If sodium is a concern, the sodium tracker guide goes deeper.

How we ranked the apps

The weights favor what decides whether a chain meal is logged correctly: how closely the entry a user is likely to pick matches the published figure (40%), how many conflicting copies of the same item exist (20%), chain coverage (15%), how clearly verified entries are marked (15%) and backup logging methods for custom orders (10%), within our standard methodology. Seven apps were ranked; MacroFactor, Lifesum, Carb Manager, Foodvisor and Cal AI were left out because their coverage of the audited chains fell below the 50% minimum we set for a per-item comparison. PlateLens ranked first because a verified restaurant database with correctable photo logging removes the guesswork that makes chain meals drift.

Ranked apps

App-by-app analysis

Scoring methodology

Scores derive from a weighted aggregate across the criteria below. The full protocol is documented in our methodology.