Icône de l'app Plateful: Calorie Tracker

Plateful: Calorie Tracker

Non revendiquée

On an iPhone running iOS 27 with Apple Intelligence, Plateful estimates your meals on…

Par Asad Ansari · Santé et forme
Score Debuts 76NouveauSois le premier à la noter
Plateful: Calorie Tracker, capture d'écran 1Plateful: Calorie Tracker, capture d'écran 2Plateful: Calorie Tracker, capture d'écran 3Plateful: Calorie Tracker, capture d'écran 4Plateful: Calorie Tracker, capture d'écran 5

En bref

Plateful: Calorie Tracker is a health and fitness app that estimates meals from photos, typed text, or speech, with on-device estimation on iOS 27 and Apple Intelligence as the default path. It is aimed at people who want food logging without sending their meal photos to a server, and it draws its numbers from a bundled database of about 12,900 USDA-based foods. A standout detail is that cloud estimation through Claude is off by default and only sends a photo after stripping its location and camera metadata.

Résumé par Debuts, d'après la fiche de l'App Store.

À propos

On an iPhone running iOS 27 with Apple Intelligence, Plateful estimates your meals on the device itself. The photograph is read where it was taken and is never uploaded.

Food logging asks for something unusually personal: what you eat, when, how much, and a picture of it, every day for months. Plateful is built so that record has nowhere to go.

HOW A MEAL IS ESTIMATED, AND WHAT EACH PATH SENDS

1. On your iPhone. On iOS 27 with Apple Intelligence, photos, typed meals, and spoken meals are read by the model already on your phone. Nothing is uploaded. No account, no monthly limit, no network needed.

2. Apple's Private Cloud Compute, at your tap. One tap on the confirm card sends a hard-to-call meal to Apple's Private Cloud Compute for a second opinion from a larger model. It runs on Apple's hardware-attested servers, Apple retains nothing, the request is not readable by Apple, and it never passes through Plateful's servers. It needs no account, and it happens only when you press the button.

3. Claude, if you switch it on. Cloud estimation is how an iPhone without Apple Intelligence gets an AI estimate, and Pro subscribers can choose it deliberately for a harder meal. Before anything is sent, the photo is downscaled and stripped of location and camera metadata on the device. Plateful's server hands the meal to the model and keeps nothing: no image storage, no image column, no log of it. Your meals are not used to train models. The switch starts off, and Settings turns it back off.

WHAT NEVER LEAVES THE PHONE, IN ANY MODE

Your corrections to an estimate. Everything read from Apple Health: weight, workouts, active energy. Your weigh-ins. Your goals. None of it syncs, none of it is sent, and none of it reaches an estimation request.

Plateful carries no advertising, no third-party analytics, and no tracking across apps or websites.

WHERE THE NUMBERS COME FROM

Privacy is worth little if the figures are guesses. The model reads what is on the plate and judges how much there is. The numbers come from a database of roughly 12,900 foods built from USDA FoodData Central, bundled in the app and needing no network. Every saved meal records the reference it was priced from.

WHAT YOU NEED

Les notes des utilisateurs

Connecte-toi pour noter cette appli. Ce qui compte, c'est d'être honnête.

Pas encore de notes. Tu l'as essayée ? Sois le premier à la noter.

Signaux de découverte

Comment l'IA et les gens découvrent Plateful: Calorie Tracker sur Debuts
Prête pour l'IAEn partie
78 % décrite pour être repérée par l'IA
Visible dès maintenantEn ligneDonnées structurées sur cette page et l'API Debuts

Dis-nous ce que tu aimes. On t'écrit seulement quand une nouveauté correspond, et les créateurs ajoutent souvent un avantage pour les testeurs.

Quel type d'apps
Sur mon
Me montrer uniquement
À quelle fréquence
On t'envoie un lien de confirmation. Rien n'arrive tant que tu n'as pas cliqué dessus.