Photo your plate. AI identifies it. Macros logged in seconds.

That’s the pitch for MyFitnessPal’s Photo Upload feature, which launched February 24, 2026 as the centerpiece of the Winter Release. We tested it across 15 meals over four days: grilled chicken and rice, a Chipotle burrito bowl, and a home-cooked Thai curry. The results are more complicated than the marketing suggests.

The bigger question: does any of this justify upgrading to Premium+ at $99.99/year?

Short answer: for most people, no. But the details matter.

Quick Verdict

Aspect	Rating
AI Photo Accuracy (simple meals)	★★★★☆
AI Photo Accuracy (complex meals)	★★☆☆☆
Blue Check Recipes	★★★☆☆
Meal Planner (new Recipes tab)	★★★☆☆
Value at $99.99/year	★★☆☆☆

Best for: Existing Premium+ subscribers who eat simple, identifiable meals and want faster logging Skip if: You track macros precisely, eat a lot of mixed dishes, or are starting fresh without existing MFP history Free tier: Calorie and macro tracking, 14-million-item food database, barcode scanner, exercise logging Premium: $79.99/year — adds meal-level macro goals, food analysis, ad-free experience, Meal Scan Premium+: $99.99/year — adds Blue Check Collection, improved Meal Planner, GLP-1 tracking

What I Actually Tested

Four days, 15 meals, iPhone 15 Pro. The test included:

• Four simple plates (chicken breast + rice + broccoli; salmon + roasted vegetables; eggs and toast; protein shake with banana)
• Four restaurant meals (Chipotle bowl, sushi platter, pizza slice, Sweetgreen salad)
• Four home-cooked meals with sauces (Thai red curry, pasta bolognese, stir fry, chili)
• Three mixed/ambiguous dishes (açaí bowl, breakfast burrito, shakshuka)

For each meal, I photographed the plate, accepted whatever the AI suggested, then manually verified by weighing and entering each component individually. This gave me a ground-truth comparison.

How the Photo Logging Actually Works

The feature is called Photo Upload, which is subtly different from the live-camera Meal Scan that existed before. You can photograph your plate and log it later rather than having to scan in real-time at the table. The AI sends the image through MFP’s database to match visual output against actual food entries.

The technology pipeline: computer vision identifies food components, cross-references them against MFP’s 14-million-item database (not generic nutritional estimates), and populates a log entry. If it sees salmon and rice, it finds “Atlantic salmon, cooked” and “white rice, cooked” in the database rather than making up numbers.

That database integration matters. It’s what separates MFP’s approach from apps that just generate a calorie estimate from the image. The entries come from identifiable database records, which means you can edit the portion size against a real entry.

The Accuracy Results

Simple meals: genuinely good. Chicken, rice, and broccoli: the AI correctly identified all three components, grabbed reasonable database entries, and landed within 8% of my manually entered total calories. For the salmon and vegetables, it nailed the fish, slightly misidentified the vegetables (called Brussels sprouts “mixed greens”), but landed within 12% on calories. For basic clean meals, this is fast and close enough for general health tracking.

Restaurant meals: inconsistent. The Chipotle bowl was the clearest win. MFP’s database has excellent Chipotle entries from the chain’s publicly disclosed nutritional data, so it matched a specific Chipotle bowl entry accurately. The sushi platter was a mess: it identified some pieces correctly but missed the rice entirely in several rolls, which is where most of the calories are. Total underestimate: 28%.

Complex home-cooked meals: where it breaks down. The Thai curry underestimated by roughly 40%. It identified the chicken and some vegetables but couldn’t see the coconut milk and fish sauce that accounted for most of the fat and sodium. The chili underestimated by 32%. If you cook with sauces, marinades, oils, or anything where ingredients disappear into a dish, photo logging will consistently miss calories.

This isn’t surprising. A peer-reviewed study published in 2024 (PMC11314244) tested seven AI food logging apps including MFP and found that even with 97% component recognition accuracy, energy estimation errors were significant—particularly for mixed and sauced dishes. One example from the study: “Eggs on toast with butter” was underestimated by 35% because butter was invisible to the AI. MFP still outperformed competitors (Fastic scored 92%, Lose It! scored 46%), but “best in class for photo logging” doesn’t mean “accurate enough to rely on.”

The honest bottom line on accuracy: If you’re tracking for general awareness and healthy habits, photo logging is fast and close enough. If you’re in a calorie deficit for body composition goals and accuracy within 5-10% matters to you, you still need to manually verify portions and cross-check entries.

What’s New in the Winter 2026 Update

Photo Upload vs. Meal Scan: The previous Meal Scan feature required you to scan in real time. Photo Upload lets you shoot the plate, close the app, and log whenever you want. This is a genuine usability improvement—logging at the dinner table is awkward and antisocial. Being able to photograph and log later is better.

Improved AI serving sizes: MFP updated the AI behind Meal Scan in January 2026 to improve portion accuracy. I didn’t have a January baseline to compare against, but the February version still struggles with portion estimation for anything more than a few inches thick or stacked.

Blue Check Collection (Premium+): This is a curated set of registered dietitian-reviewed recipes that Premium+ subscribers can access. The goal is to distinguish verified nutrition data from the user-submitted entries that make up most of the database. In practice, the selection is limited—this is a small curated library, not a replacement for the broader database. It’s useful for meal planning if you actually use MFP’s recipe features; irrelevant if you don’t cook from the app.

Improved Meal Planner with Recipes tab (Premium+): The Meal Planner now has a dedicated Recipes tab that saves favorites for future planning. The interface is cleaner than before. For people who plan meals a week out and grocery shop from the app, this is a legitimate quality-of-life improvement. For people who track meals retroactively (which is how most people use MFP), it’s irrelevant.

Cameron Brink partnership: WNBA player Cameron Brink worked with MFP on the Winter 2026 release, sharing her experience using the app to track protein during injury recovery while working with a dietitian. This is a marketing collaboration, not a product feature. The partnership signals where MFP is heading: performance athletes and GLP-1 users are the growth segments, a shift away from their traditional general weight-loss user base.

Free vs. Premium vs. Premium+: Where the Wall Really Hits

Free tier gets you:

• Full 14-million-item food database
• Calorie and macro tracking
• Barcode scanner
• Exercise logging
• Basic nutrition dashboard

Premium ($79.99/year) adds:

• Meal-level macro goals (set carb/protein/fat targets per meal, not just daily)
• Detailed food analysis and micronutrient breakdown
• Calorie goal adjustments by exercise intensity
• Ad-free experience
• Meal Scan and Photo Upload AI logging

Premium+ ($99.99/year) adds over Premium:

• Blue Check Collection (RD-reviewed recipes)
• Improved Meal Planner with Recipes tab
• GLP-1 medication tracking (dose, timing, injection site)
• Expanded nutrition reports

The $20/year jump from Premium to Premium+ gets you the recipe library, better meal planning, and GLP-1 tracking. Photo logging is included in Premium, not Premium+. If photo logging is your primary reason to upgrade, you only need Premium at $79.99/year.

This is worth clarifying because MFP’s marketing has bundled the photo logging announcement with Premium+ news, which creates confusion about which tier the feature requires. You don’t need $99.99/year for AI photo logging. You need $79.99/year.

Premium+ vs. the Alternatives

At $99.99/year, Premium+ is competing directly with MacroFactor ($71.99/year) and Cronometer Gold ($49.99/year). The comparison isn’t flattering for MFP.

MacroFactor’s nutrition app uses the NCC Food and Nutrient Database (a research-grade source) and adapts your calorie targets weekly based on your actual weight trend. If you’re in a calorie deficit and your weight plateaus, MacroFactor tells you to eat less. MFP tells you nothing—you have to recalculate manually. For body composition work, that adaptive intelligence is more valuable than photo logging.

Cronometer Gold ($49.99/year) has no photo logging, but its database is derived from authoritative sources (USDA, manufacturer labels, peer-reviewed composition data). If micronutrient tracking matters to you—and it should if you care about long-term health, not just macros—Cronometer’s verified entries are more reliable than MFP’s crowdsourced ones.

Where MFP still wins: database breadth. 14 million foods means the specific cereal you eat, the regional chain restaurant near your office, and the exact protein bar in your gym bag are probably in there. MacroFactor’s database is smaller. Cronometer’s is smaller. For people who’ve spent years building logging habits around MFP’s database, switching has real costs.

Who Should Actually Upgrade

Pay for Premium ($79.99/year) if: You’re an existing MFP user who logs consistently and wants photo logging to reduce friction. Simple, identifiable meals logged faster is worth $80/year for regular users. The meal-level macro goals are also genuinely useful if you’re eating around specific workout times.

Pay for Premium+ ($99.99/year) if: You’re on a GLP-1 medication (Ozempic, Wegovy, Mounjaro, Zepbound) and want purpose-built dose and timing tracking. Or you meal plan proactively and want verified recipes from registered dietitians. The extra $20/year is justified for these specific use cases; it’s not justified just for photo logging.

Stay on free if: You track calories loosely and aren’t obsessing over macro precision. The free tier has enough capability for general tracking with the full food database, barcode scanner, and exercise log.

Look elsewhere if: You’re evaluating nutrition apps from scratch. Starting with a $100/year commitment to MFP when you have no existing data history is hard to justify. MacroFactor’s 7-day free trial (no card required) lets you evaluate genuinely adaptive nutrition coaching before committing. And if photo logging is your primary requirement, apps like Lose It! offer similar features at lower price points.

The Database Problem Photo Logging Doesn’t Fix

Here’s the issue that no AI photo feature solves: MFP’s crowdsourced database has accuracy problems. User-submitted entries include duplicates, outdated restaurant nutritional data, and incorrect macros on common foods. The Blue Check Collection addresses this for a small curated subset of recipes. The broader database has the same reliability problems it’s had for years.

When the AI photo logger cross-references your meal against the database, it pulls from that broader database. A more accurate photo AI matched to an unreliable database still produces unreliable output. This is a structural issue with MFP’s approach.

For people doing serious body composition work—precision calorie deficits, exact macro splits, body recomposition—this matters. For general health tracking, it probably doesn’t.

The AI Fitness App Context

MFP’s photo logging update is part of a broader shift in fitness apps toward AI-assisted data entry. The best AI fitness coach apps in 2026 are increasingly trying to reduce the manual data entry burden, because logging friction is the number one reason people quit nutrition tracking. MFP’s Photo Upload addresses that exact problem.

The question is whether reducing logging friction is worth the price jump. For someone who has repeatedly tried manual food logging and quit because it felt like a part-time job, the answer is yes—provided your diet consists mostly of simple, identifiable meals. For athletes and performance-focused users who need macro precision, photo logging is a supplementary tool, not a replacement for careful entry. Check our guide to fitness apps without the obsession spiral if you’ve found nutrition tracking apps do more harm than good—photo logging that produces fuzzy estimates can actually reduce obsessive rechecking for some people, which is a legitimate benefit.

The Verdict

Photo Upload works. For simple meals, it’s fast, reasonably accurate, and meaningfully reduces logging time. For complex meals with sauces, layered ingredients, or anything that isn’t visually distinct, it misses enough calories to matter.

At $79.99/year (Premium), AI photo logging is a reasonable value for existing MFP users who log consistently. At $99.99/year (Premium+), the additional features (RD recipes, improved meal planner) are marginal unless GLP-1 tracking applies to you.

For fresh evaluators: start with the free tier for two weeks. If you’re logging regularly and want the photo feature, Premium at $79.99/year makes sense. Premium+ is a $20 upcharge for features most people won’t use.

For existing Premium subscribers: the Winter 2026 update is a solid incremental improvement. Photo Upload is the best new feature; the improved Meal Planner is secondary. The AI is getting better at recognizing simple foods. It’s not replacing manual logging for precision tracking.

Download MyFitnessPal on iOS or Android. Photo Upload is available now for iOS Premium and Premium+ subscribers. Android is on the waitlist with no committed timeline.

---

Tested February 25-28, 2026 on iPhone 15 Pro. 15 meals photographed and compared against manually weighed and entered values. Free tier accuracy testing conducted in parallel for database comparison.