Research-Grade
Nutrition Tracking
1.5 seconds per meal. 1.6 minutes per week. 95% accurate. Learns your recipes.
Manual Logging Is Killing Your Progress
80% of people quit tracking nutrition within 4 weeks. Here's why:
Manual logging takes forever
42-63 minutes per week searching databases, adjusting portions, double-checking totals.
Generic databases are inaccurate
Generic food databases don't know your portions. ~30% error on homemade meals.
No learning, no improvement
Same tedious process forever. 80% quit within 4 weeks. People give up.
How It Actually Works: 1.6 Minutes Per Week
Not magic. Just physics, statistics, and network effects.
3 Weigh-Ins Per Meal Type
Weigh your chicken breast 3 times. Your curry 3 times. Your oatmeal 3 times. System learns your Gaussian portion distribution (mean Β± std dev).
Chicken breast: 195g, 210g, 200g
β Mean: 202g Β± 8g
β Calibrated forever
90% of Meals: 1 Second
Open app β See top 10 predicted meals β Tap "Chicken & Rice" β Done. System applies your calibrated 202g Β± 8g portion automatically.
β’ 90%: Single tap (1 sec)
β’ 9%: Quick search (3 sec)
β’ 1%: Weigh novel item (30 sec)
= 1.5 sec/meal
Expiry-Based Physics
Milk expires in 7 days β System auto-reconciles all meals with milk β Learns your actual usage β 95% accuracy without manual work.
β’ Receipt scan: 5 sec
β’ 21 meals Γ 1.5 sec: 31 sec
β’ Reweigh 5 staples: 60 sec
= 96 sec (1.6 min)
Competitors: 42-63 min/week β’ Eatomate: 1.6 min/week
Save 44 hours per year. That's over a work week of your life back.
What You Get From Every Meal Log
Each 1.5-second meal log gives you complete nutritional data
π Instant Meal Data
- Meal weight in grams
Precise portion size for every meal
- Total calories (kcal)
Accurate calorie count for this exact meal
- Macronutrients breakdown
Protein, carbs, and fat in grams
- Micronutrients & vitamins
Vitamin A, C, D, B12, calcium, iron, and more
- 100% meal identity confirmation
Accurate identification of exactly what you're eating
π Available Tracking Tools
- Daily nutrition tracking
See daily totals for calories and all nutrients
- Complete meal history
Full timeline of all meals with nutrition data
- Pantry inventory management
Track groceries with barcode and receipt scanning
- Personalized recipe learning
System learns your cooking patterns over time
- Accuracy improvement tracking
Watch your accuracy rise to 95%+ in 4 weeks
- Goal setting & progress
Set nutrition targets and track against daily goals
How Eatomate Is Different
95% Accuracy Through Physics
Physics-based reconciliation eliminates systematic errors. Weekly nutrition reports reach 95% accuracy (95% CI, 2Ο).
Learn about our accuracy system βLearns your Recipes (Personalized)
Not generic database approximations. Learns your grandmother's sambar (8.5g oil/100ml, not generic 3g). Your bolognese (12g oil/100g, not generic 5g). Your cooking style.
Reconciliation learns ingredient ratios unique to your household automatically.
5-Second Receipt Scan (Network Effects)
Scan receipt β Mistral AI auto-maps items β Done in 5 seconds. First user maps "Tesco Milk 2L" β Saved for all users forever. By month 12: 95%+ cache hit rate.
2M+ products preloaded. Zero manual mapping needed.
Gets Smarter Over Time
Weekly accuracy improves to 95%+ in just 4 weeks. Four learning layers improve continuously.
See the learning timeline βBacked by Science
Research-grade accuracy through scientifically validated methodology
95% Accuracy vs reality in 4 weeks
You search, weigh, and log each meal item. AI matches your entries to our database. Physics learns your recipes through pantry reconciliation.
Includes restaurant meals, brand variation, and database errors.
Weekly accuracy after 4 weeks using mass conservation physics
Barcodes preloaded for instant nutrition lookup
Recipes covering global cuisines
Eatomate uses physics-based reconciliationβa novel approach to nutrition tracking that eliminates systematic errors.
Read the science behind Eatomate βWhy It Works This Way
The engineering story behind 1.6 minutes per week
Before settling on what Eatomate is today, I tried everything. 3D reconstruction from photos using COLMAP. Depth-based scanning. Fine-tuned AI vision models trained on food images. Neural Radiance Fields (NeRF) to build detailed 3D food models from 2D photos. 3D Gaussian Splatting. I even seriously considered acoustic spectroscopy using a smartphone speaker β before coming to my senses. And yes, at one particularly desperate 2 AM, I genuinely explored analysing heat distortion patterns in steam rising from food to infer its contents. That one I'm not proud of. Each approach was more technically impressive than the last.
Every single one was less accurate than a Β£10 kitchen scale. Not slightly less β significantly less. The fundamental problem with visual estimation is that density varies wildly. A bowl of rice and a bowl of risotto look identical to a camera. Volume tells you almost nothing about mass, and mass is what your body actually processes.
So I stopped trying to be clever. A kitchen scale gives you ground truth in grams. Mass conservation physics means when your milk carton empties, the system knows exactly how much went into every meal that week. Simple inputs. Rigorous maths. That's why it works.
Sometimes the boring solution is the right one.
Start Tracking the Right Way
1.5 seconds per meal. 1.6 minutes per week. 95% accurate.
Start your free 14-day trial today.
No credit card required. Cancel anytime.