Snap a meal.
Done.

Photo-based meal logging. The AI reads the plate, estimates calories and macros, and the coach uses it to shape the rest of your day.

Join the waitlist Show me how
iPhone and Android · free in beta
How it works

Two seconds,
not two minutes.

Take the photo before you eat, or while you eat. Most meals are logged in under two seconds. The path goes like this.

1

Snap or import

Take a photo, or pick one from your camera roll. Works for plated meals, takeout, and most things on a fork.

2

AI reads the plate

Calories and macros come back with a confidence score. Around 90% on common meals. Lower if it is something it has not seen often.

3

Adjust if needed

Sauce-heavy bowls and mixed dishes sometimes need a tap. Swap an ingredient or fix the portion in one tap.

4

Logged, coach sees it

The day rolls up automatically. The coach reads it for the rest of its advice that day.

Chicken burrito bowl with rice, beans, and toppings
82% sure
chicken burrito bowl
640 kcal · 52g protein
Berry baked oatmeal in a baking dish
94% sure
oat porridge + berries
380 kcal · 14g protein
The other 10%

Type it when photos
won't cut it.

Photos handle most home and restaurant meals. Some things don't photograph well: a beer at the bar, a packaged protein bar, a sauce-heavy curry. Those have other ways in.

Barcode scan

Same camera, different mode. Reads the barcode and pulls from a verified packaged-food database.

Text search

Type 'chipotle bowl 720 cal' or just 'two beers'. Pulls from a large food database with most chains and packaged foods.

Recent meals

Tap from yesterday or last week. One tap, logged. Most people eat the same eight meals on rotation anyway.

Feeds the coach

Every meal makes
the coach sharper.

Once a few days of meals are in, the advice gets specific. Calories, protein, and timing all show up in what the coach recommends. Especially when training and eating disagree.

Protein gap
Signal

"Under 140 g protein three days running."

Coach

Mid-afternoon nudge: 'You've got a protein gap. Greek yogurt or a chicken wrap before the gym?'

Deficit during a strength block
Signal

"800 calories under target during a heavy lifting week."

Coach

When bench felt heavy yesterday, the coach explains why. Suggests easing the cut for the next 4 days.

Cross-context
Signal

"Calories under maintenance plus 5 hours of sleep."

Coach

Coach sees both, recommends a rest day instead of pushing the planned PR attempt.

FAQ

Photos, macros,
honest answers.

Have a meal we should test on? Send it →

Around 89 to 94% on common meals (versus weighed-and-logged), based on beta testing. Confidence shows on every photo. If it is under 70%, fix it before saving.

Mixed dishes, soups, and casseroles are the hardest. The app shows what it thinks the food is. If wrong, swap it for the right entry from the database in one tap.

Restaurant meals: usually yes for plated foods, less so for sauce-heavy stuff (curries, stir-fries). Packaged foods: barcode scan is faster and more accurate. Both routes are in the app.

Yes. Same camera, different mode. Most packaged foods are in the database, plus verified entries from the user community.

Yes. Calories, protein, carbs, and timing all feed into what it suggests. Hard training week with low calories means a different recommendation than easy week with good food. Less logged data means more general advice.

Stop tracking.
Start training.

Join the waitlist Talk to us
Join the waitlist