new · snap a meal~2 seconds

AI photo
calorie counter.One snap.

Point a camera at a plate. The AI identifies the food, estimates portion size from plate-edge detection, and returns calories and macros in under 3 seconds. No database scroll. No 12 variants of “chicken sandwich.”
  • Within 8% of a registered dietitian
  • Works on restaurant photos + home plates
  • One-tap adjust if a portion looks off
  • Free on the web, no signup to try
Try the AI · live
Try the photo flow.
Snap or upload a meal photo. The AI returns calories and macros in seconds. Or type if you'd rather.
today
0 / 2,000 kcal
Photo to calories in 3 steps

How a photo becomes a number.

01 · Detect
Detect
A vision model identifies the foods on the plate. "Grilled chicken, white rice, broccoli."
02 · Estimate portion
Estimate portion
Plate scale, utensil cues, and visual volume become grams per food. The hardest step.
03 · Look up nutrition
Look up nutrition
Grams × USDA FoodData Central values. Plain arithmetic, not a guess.
Photo tips

What makes a good photo.

You don't need food-blog lighting. You need a photo that lets the model see the food.

do
Plate centered, top-down or 45°
avoid
Side-on shot that hides depth
do
Most foods visible (stir if needed)
avoid
Everything hidden under sauce
do
One frame, one meal
avoid
Breakfast + lunch + coffee in one shot
do
Daylight or normal room light
avoid
Total darkness or hard backlight
Strengths and weaknesses

Where the photo flow wins.

Wins
  • · Restaurant meals at known chains
  • · Home stir-fries, pastas, salads, rice bowls
  • · Sharing plates (snap before eating, adjust to half)
  • · Snacks small enough that you'd normally skip logging
Less reliable
  • · Mixed soups and stews (hidden ingredients)
  • · Buffet plates with many small items
  • · Foods where prep varies a lot (fried chicken oil)

For these, type the meal name and add a one-line detail. The text flow handles it well.

AI photo calorie counter · FAQ

Questions, answered.

Our benchmark target is within 8% of a registered dietitian on average, with median error under 5% on common meals. Peer-reviewed research puts mainstream AI photo calorie estimation in the 5 to 15% range, which is roughly the same band as a dietitian eyeballing the same plate.

One photo. Two seconds. Done.

The fastest way to log a meal you didn't cook yourself.