Why Is the AI Giving Wrong Calorie Estimates?

Nutrify's AI estimates calories from photos and descriptions, which means accuracy depends on what the AI can see or understand. This article explains common reasons for inaccurate estimates and how to get better results.

Why AI Estimates Can Be Off

The AI makes a visual or textual guess based on what it observes. It cannot weigh your food. Accuracy is affected by:

• Portion size - A blurry or distant photo makes it hard to judge how much food is on the plate
• Lighting - Dark or overexposed photos hide details the AI uses to identify food
• Plate angle - Overhead shots work better than angled shots for estimating volume
• Multiple items - A crowded plate with overlapping foods is harder to analyse accurately
• Mixed dishes - Stews, curries, and casseroles have variable ingredient ratios that are difficult to estimate visually

Tips for More Accurate Scans

1. Take the photo from directly above - A top-down view gives the AI the best view of portion size
2. Use good lighting - Natural daylight or a bright room works best. Avoid shadows across the food
3. Keep the food in frame - Make sure the whole meal is visible, not cut off at the edges
4. Scan individual items separately - For mixed plates, scan each component separately and add them together
5. Use the description field - If the photo is ambiguous, tap the text input and describe the food and portion (e.g. "2 scrambled eggs with 2 slices of toast")
6. Specify weight when you know it - Adding "200g chicken breast" in the description gives the AI a concrete anchor

💡 Describing your food with portion size ("a large bowl" vs "a small bowl") significantly improves accuracy even with a good photo.

Editing a Result

If the AI estimate looks wrong, tap Edit on the result to adjust calories and macros manually before logging. Your correction is applied to that entry immediately.

Reporting a Bad Estimate

We use feedback to improve the AI. To report a wrong estimate:

1. After receiving a result, tap the thumbs down or feedback icon
2. Describe what was wrong (e.g. "calories too high, I had 100g not 200g")
3. Submit the report

Our team reviews flagged results regularly.

Known Limitations

• Liquid-heavy foods (soups, smoothies) are harder to estimate accurately than solid foods
• Homemade or restaurant dishes vary widely from standard database entries
• Multi-ethnic and regional foods may have fewer training examples, leading to wider estimates

For verified nutritional data, use the barcode scanner or search the food database by name to find an exact match.

Need More Help?

If you have a question not covered here, the fastest way to reach us is directly inside the Nutrify app:

1. Open Nutrify
2. Go to Settings
3. Tap Help & Feedback
4. Tap Report a Bug or Request a Feature

Our team reads every message. You can also email us at support@nutrifytracker.com.