Food tracking: analyze a meal and follow your nutritional needs

Weighing every ingredient and looking up its nutrition facts discourages most people within three days. Happy Assistant tackles the problem from the other end: a photo of your plate, or a sentence describing the meal, and you get an estimate of calories and macronutrients, plus an alert if the dish conflicts with your saved restrictions.

Key takeaways

A photo or a description is enough: no scale, no barcode, no database to search.
Estimates for calories, protein, carbs, sugar, fat, fiber and sodium.
Automatic alert if a food conflicts with your allergies or diet.
Every meal is logged in a daily journal with running totals.
Personalized daily needs, computed from your gender, age, weight, height and activity level.
Your real workouts change your targets: measured activity beats what you declared.
Meals sorted automatically by time slot (breakfast, lunch, dinner, snack) and correctable by voice.

Phrases that work

Analyze this photo of my lunch.
For lunch I had a salmon, avocado and rice poke bowl.
How many calories have I eaten today?
Remove this morning's breakfast from the log.
Remember that I weigh 71 kg, am 1.80 m tall and work out three times a week.
Actually there were also two slices of bread with that lunch.

How an analysis works

Open the assistant and attach a photo of your plate (or describe the meal by voice).
The assistant identifies the visible foods and estimates the portions.
It cross-references the content with your food memory: allergies, intolerances, diet.
It announces the calories and macros, flags points of attention, then logs the meal.
The day's total is updated in Profile → My diet.

What the food journal contains

The day's meals, sorted by type (breakfast, lunch, dinner, snack).
The day's running totals, compared with reference benchmarks.
The history of the last seven days, to spot a trend rather than a snapshot.
The detail of each meal: identified foods, macros, alerts, confidence level.

The point of the journal isn't gram-level precision, but regularity: logging your meals for a few days is enough to spot an imbalance — too little fiber, too much sodium, protein concentrated in a single meal.

Nutritional needs computed from your profile

Comparing your intake with “2,000 kcal” teaches you nothing: that benchmark matches nobody in particular. Happy Assistant computes your needs with the Mifflin-St Jeor formula, from the gender and date of birth in your profile, then the weight, height and activity level you shared with it. The bars on the Analysis screen and the assistant's comments rely on those targets.

The declared activity level is only a starting point: as soon as you log workouts, the actually measured expenditure takes over. The app averages your training over the last fourteen days — and includes already planned sessions — to adjust your calorie needs. Taking up sport? Your targets go up without you re-entering anything.

Same rigor for protein: the target is not a single number, but a coefficient in grams per kilogram of body weight chosen according to your actual practice — sedentary, endurance, strength, mixed — and raised if you are in a calorie deficit, to protect muscle mass. A runner and a weightlifter don't have the same needs, and the app knows it.

Calories, protein, carbs, fat, fiber and salt adjusted to your estimated metabolism.
Nothing to fill in a form: say “I weigh 71 kg and I'm 1.80 m”, and the information is remembered.
Recalculation is immediate: as soon as weight, height or activity changes, the target changes.
While weight or height is missing, the app shows generic adult benchmarks and tells you so.

A journal that requires no discipline

Each meal is filed automatically in its time slot: breakfast, lunch, dinner, snack.
Foods reported across the same meal (a window of about 3 hours) are grouped into a single entry.
A spoken correction — portion, forgotten food, wrong slot — updates the already saved meal.
The journal lives in Profile → My diet, with the last 7 days so you can read a trend.

Alerts and dietary restrictions

If your memory contains “tree nut allergy,” a pine nut pesto or a cashew curry triggers an explicit alert before any other comment. Diets (vegetarian, gluten-free, lactose-free) are handled the same way.

An alert isn't a certainty: it invites you to check the dish's actual composition, especially at a restaurant.

Method, uncertainty and sources

A photo identifies foods and estimates portions; nutritional values are then matched to reference food data. Oils, sauces and hidden ingredients remain the main uncertainty. Correcting a portion or ingredient recalculates the meal.

Measurement: quantities you explicitly provide.
Estimate: visible portions, calories and nutrients inferred from a photo.
Recommendation: comparison with personalised benchmarks, never a diagnosis.
References: Mifflin-St Jeor resting-energy equation; ANSES and EFSA nutrition guidance.

What the feature does not do

These are estimates: a photo reveals neither the cooking oil nor the hidden sugar in a sauce.
The assistant states its confidence level; specify the portion to refine the calculation.
This is not medical advice: for a therapeutic diet, rely on a healthcare professional.

Frequently asked questions

Can the assistant remember my usual meals?

Yes. You can have it memorise a usual meal (“remember my usual breakfast: 40 g of oats, a yoghurt, a banana”), a product's nutrition label (values per 100 g) and your typical portions (“my bowl of rice is 120 g”). After that, “the usual” is enough: the meal is logged with the right quantities, with no visual re-estimation.

Do my needs change if I exercise more?

Yes, automatically. As soon as sessions are logged in your training journal, the app averages your real expenditure over fourteen days (planned sessions included) and uses it instead of the declared activity level. The protein target adapts too: it is higher for strength training or during a calorie deficit. Nothing to re-enter.