Skip to main content
The food scan feature lets athletes photograph a meal and instantly see its estimated nutritional content. Point the camera at any plate of food, and Astral’s AI identifies what’s in the image and returns an estimate of calories, protein, carbohydrates, fat, and fiber. There’s no manual logging — just a photo.
Food scan is currently in early access. The feature is enabled on a per-account basis. If you don’t see it in the app, reach out to your Astral coach or account contact to request access.

​
How to use food scan

1

Open the nutrition tab

In the Astral app, navigate to Nutrition from the bottom navigation bar.
2

Tap Scan food

Tap the Scan food button (camera icon) to open the camera.
3

Photograph your meal

Point the camera at your food so the full plate is visible. Good lighting and a clear view of all items will give you the most accurate result.
4

Review the result

The app displays the AI’s nutrition estimate within a few seconds, broken down by item and totalled across the whole meal.

​
What the result includes

For each scan, you receive:
  • Food name — a short description of what the AI identified (e.g., “Grilled chicken salad”)
  • Calories — total estimated kilocalories
  • Protein — grams of protein
  • Carbohydrates — grams of carbs
  • Fat — grams of fat
  • Fiber — grams of fiber
  • Individual items — a breakdown per food item visible in the image, each with its own calorie count and estimated portion
  • Confidence level — high, medium, or low, indicating how clearly the AI could identify the food

​
Example result

{
  "food_name": "Grilled chicken salad with avocado",
  "calories": 520,
  "protein_g": 42,
  "carbs_g": 18,
  "fat_g": 28,
  "fiber_g": 7,
  "confidence": "high",
  "items": [
    { "name": "Grilled chicken breast", "calories": 280, "portion": "150g" },
    { "name": "Mixed greens", "calories": 20, "portion": "2 cups" },
    { "name": "Avocado", "calories": 160, "portion": "half" },
    { "name": "Cherry tomatoes", "calories": 25, "portion": "6 pieces" },
    { "name": "Olive oil dressing", "calories": 35, "portion": "1 tbsp" }
  ]
}

​
Accuracy and confidence

Food scan estimates are based on what’s visible in the photo. A few things affect accuracy:
  • Portion size — estimates are based on what the AI can see. Unusual portion sizes or partially hidden food may be off.
  • Mixed dishes — soups, stews, and casseroles are harder to break down item by item. Expect a medium or low confidence rating for these.
  • Image clarity — blurry images, low light, or photos taken at an angle reduce accuracy. A direct overhead or slight-angle shot in good light works best.
When the AI can’t confidently identify the food, it returns a low confidence rating. This is a signal to double-check the result before logging it.
If a result comes back with low confidence, try retaking the photo in better light or from directly above the plate. Spreading out mixed items (like a stir-fry) so individual components are visible can also improve accuracy.

​
Image requirements

  • Accepted formats: JPEG, PNG, and other common image types
  • Maximum file size: 5 MB
  • The image must be sent as a base64-encoded string

​
Non-food images

If you photograph something that isn’t food, the AI returns a result with "food_name": "Not food" and zeroed-out nutrition values rather than an error. The app will display a message letting you know to try again with a food photo.

​
Availability

Food scan is gated behind early access and is not available to all accounts by default. Access is enabled per user by the Astral team.