How Forge AI estimates calories and macros.
Forge AI is designed for honest, useful food estimates. It does not pretend to know exact calories when the input is vague. Instead, it uses the best available signal: source-backed restaurant data when available, barcode lookups for packaged foods, and AI estimation for real-world meals that do not have a simple label.
Try the food loggerThe estimation workflow
1. Parse the meal
Forge looks for food items, portions, drinks, sauces, sides, brands, restaurant names, and phrases like "half" or "light dressing."
2. Use sources when possible
Known chain items and packaged foods can use source-backed nutrition or barcode lookup instead of a generic estimate.
3. Estimate the rest
For homemade and custom meals, AI estimates calories, protein, carbs, and fat from food composition and the user description.
What improves accuracy
The best descriptions include the calorie-driving details: portion size, cooking method, oil, sauce, dressing, cheese, nuts, breading, drinks, sides, and whether the user finished the meal. "Romaine salad with red onion, light blue cheese, and breaded chicken" is much better than "salad with chicken."
Where estimates can be wrong
Food estimation has uncertainty. Restaurant portions vary. Homemade meals vary. Hidden oil can add calories. A "serving" can mean different things to different people. That is why Forge AI supports corrections, portion adjustments, barcode scanning, and source-backed data where available.
The product standard
The goal is not fake precision. The goal is a food log users trust enough to use every day: fast, transparent, specific, correctable, and much less annoying than database search.
What Forge AI does not claim
Forge AI is not medical advice, does not diagnose nutrition needs, and does not claim that every estimate is exact. If a user needs clinical nutrition tracking, allergy safety, or therapeutic diet management, they should use appropriate professional guidance and tools built for that purpose.
For fitness tracking, the practical goal is consistency plus honest adjustment. A useful estimate today, corrected when needed, is usually better than a perfect log that never gets entered.