MyNetDiary

Calorie trackers all look the same on a screenshot, and they are not the same at all. The difference sits somewhere nobody photographs: the database.

A tracker built on user-submitted entries will happily tell you a chicken breast contains 90 calories or 310, because somebody typed both. Verified databases are the exception, and Cronometer is the other well-known one. You log diligently for three months, the numbers are wrong in an unknowable direction, and the entire exercise measured nothing.

MyNetDiary's central claim is that it solved this by being old and by employing people. Two million foods, verified by staff, drawn from USDA and research-grade nutrient data, with thousands of entries updated every day. This review covers whether that depth is worth choosing a tracker for, what the AI photo logging can and cannot do, and where a calorie count stops being useful information.

MyNetDiary: The Verified Numbers

Photograph The Plate, Check The Numbers

The AI feature people will try first lets you photograph a meal and have foods identified, portions estimated and the entry logged with full nutritional detail.

It is genuinely useful and it needs to be understood correctly. Recognising that a plate contains rice, grilled chicken and broccoli is something models now do reliably. Estimating that there are 180 grams of rice rather than 240 is a fundamentally harder problem, and it is the number that determines the calorie count.

Portion estimation from a two-dimensional image involves inferring depth, density and what is hidden underneath. A photograph cannot see that the rice is piled rather than spread, or that the chicken was cooked in three tablespoons of oil. Those are not marginal differences — they are the difference between a 450-calorie meal and a 700-calorie one.

The right way to use it is as a logging accelerator rather than a measuring device. It removes the friction that causes people to stop logging altogether, which matters enormously, because an approximate log you maintain beats a precise log you abandon in week two.

For anything you eat regularly, log it accurately once — weighed, or scanned from the packet — and reuse that entry. Save the photograph for restaurant meals and unfamiliar food, where an estimate is the only option available anyway.

A photograph can tell you what is on the plate. It cannot tell you how much, and how much is the entire number.

The Barcode Is Still The Most Accurate Route

For packaged food, scanning the barcode remains substantially more reliable than any AI feature, and it is free on every tier.

The reason is straightforward: a barcode resolves to a specific product with a manufacturer-declared nutrition panel. There is no estimation involved. The only remaining variable is your portion, and that you can weigh.

This creates a practical hierarchy worth internalising. Scan the barcode where one exists. Weigh and use a verified database entry where one does not. Photograph only when neither is possible — a restaurant meal, someone else's cooking, food without packaging.

Following that order costs almost no extra effort and produces a log that is meaningfully more accurate than one built on photographs alone.

GLP-1 Support Is A Genuine Repositioning

MyNetDiary has built dedicated features for people taking GLP-1 receptor agonists: medication and injection tracking, digestive symptom logging, protein targets, specialised meal plans and trend charts.

This is a serious response to a real shift rather than a marketing label. These medications changed the problem calorie trackers were built for. When appetite is pharmacologically suppressed, eating too much stops being the difficulty and new ones appear: eating enough protein to limit muscle loss, managing nausea and digestive side effects, and maintaining nutrient adequacy on substantially reduced intake.

A tracker designed around restriction is the wrong tool for that, and neither is a meal planner such as Eat This Much. Protein targets, symptom logging and meal plans built for smaller volumes are the right ones, and building them is a considerably better response than adding the letters GLP-1 to a marketing page.

The essential caveat belongs here plainly. These are prescription medications with real side effects, and nutrition while taking them is a matter for the prescribing clinician and ideally a dietitian. A tracking app is a useful instrument for following medical advice; it is not a source of it, and the site's own material states that the information it provides does not constitute medical advice.

Used as an instrument — recording protein intake, noting which meals precede symptoms, giving your clinician real data rather than recollection — it is genuinely valuable.

• Log protein first; on suppressed appetite it is the nutrient most likely to fall short.
• Record digestive symptoms alongside meals, since the pattern is what a clinician can act on.
• Bring the trend charts to appointments rather than trying to recall a fortnight of eating.
• Treat any target the app sets as a starting point to discuss, not a prescription.
• Watch for a downward trend in overall nutrient adequacy, not only in calories.

The Free Tier Is Unusually Complete

The free version includes barcode scanning, macro tracking, water logging, exercise tracking, full nutrient detail on entries, and community access — with no advertising and no account required to start.

That is a materially better free offering than this category normally provides, and it is worth being specific about why it matters. In most trackers, the barcode scanner is the first thing put behind a subscription, because it is the feature people use constantly. Giving it away means the core loop of accurate logging works without paying.

No advertising deserves mention too. An app you open five times a day, in which advertising would be shown against food and weight, is one where the absence of ads is a real quality-of-life difference and not merely a preference.

The paid tiers add advanced reports and the AI features. That is a defensible split: the AI tools cost the vendor money per use, and the free tier retains what you need for a genuinely accurate log.

It also means you can evaluate the thing that matters — whether the database contains the foods you actually eat — before paying anything. That is the only test that predicts whether you will still be using it in three months.

What A Calorie Tracker Cannot Know

A section this honest belongs in every review in this category, because the interface implies a precision the underlying science does not support.

Nutrition labels carry permitted tolerances, which in many jurisdictions run to around 20 per cent. Two identical apples differ in size and sugar content. Cooking changes availability of energy — the same potato yields differently depending on preparation. And energy expenditure varies between people with the same measurements.

So a daily total displayed as 1,847 calories carries a false precision. The honest reading is somewhere in the region of eighteen hundred, plus or minus a meaningful margin.

This does not make tracking useless — quite the opposite. Its value was never absolute accuracy; it is consistency and awareness — the same reason a training log such as SHRED works. Logged the same way over weeks, the trend is reliable even when each day's figure is not, and the act of recording changes eating behaviour on its own.

The practical implication is to weight the weekly average heavily and the daily total barely at all, and to resist adjusting your eating in response to a single day's number. That is also the mindset least likely to develop into an unhealthy preoccupation, which is a real risk with detailed food logging and one worth naming.

Two Weeks Of Honest Logging

Nutrition trackers are chosen on features and abandoned over friction. This sequence tests the friction, which is the thing that actually decides.

The bar is step two. If the foods you eat most are missing or wrong in the database, no amount of AI on top will fix it.

1. Before anything else, search the database for the ten foods you eat most often — including anything regional, home-cooked or from a local supermarket.
2. Check the entries against the packets in your kitchen. A database that is wrong on your staples is wrong on everything you will log.
3. Log a full week on the free tier only, using barcodes where possible and weighing where not. This establishes what accurate looks like for you.
4. Now photograph a few meals you have already logged accurately, and compare the AI estimate with the known figure. That difference is your error margin for every photographed meal.
5. Connect your wearable and confirm the exercise data flows without duplicating entries, which is a common and quietly distorting problem.
6. At day fourteen, look at the weekly averages rather than any daily total, and ask whether logging has become automatic or is still an effort you are forcing.

Freemium Done The Right Way Round

The model is freemium: a substantial free tier, with Premium and Premium Plus adding advanced reports and the AI tools including photo logging, the coach and menu scanning.

Two things make this structure better than the category norm. The free tier keeps barcode scanning and full nutrient detail, so accurate logging does not require payment. And the paid features are the ones with genuine per-use cost to the vendor, which is a more honest basis for a paywall than gating a core function to force conversion.

The decision worth making deliberately is whether you need the AI layer at all. If most of what you eat is packaged or home-cooked from known ingredients, barcode scanning and weighing will serve you better and cost nothing. If you eat out frequently, photo logging and menu scanning are what make logging survivable.

We do not publish amounts, since tiers change. Do check whether the tier you want is billed monthly or annually and how cancellation works before subscribing, and note the related product PlateAI from the same team if an AI-first approach appeals more than a database-first one.

Twenty Years Of Food Data

• The database is the product — over two million staff-verified foods from USDA and NCC research data, carrying up to 108 nutrients per entry, with thousands of foods updated daily.
• Staff-verified, not crowd-sourced, which is the distinction that determines whether your logged numbers mean anything at all.
• AI photo logging and menu scanning sit on top: photograph a plate for an estimate, or scan a restaurant menu for tailored suggestions.
• GLP-1 support is built in, with medication and injection tracking, digestive symptom logging, protein targets and matched meal plans.
• The free tier is genuinely complete: barcode scanning, macro tracking, full nutrient detail and exercise logging, with no ads and no account needed to start.