Documentation
Everything you need to know about using Eatomate—from 3 weigh-ins to single-tap logging.
1.5 seconds per meal • 1.6 minutes per week • 95% accuracy via physics-based reconciliation
1.6 min
Per Week Total
1.5 sec
Per Meal Average
~95%
Accuracy by Week 4
100×
Faster Than Competitors
Getting Started
Quick setup with progressive onboarding
- Progressive Onboarding (2-5 minutes)→Scan a few pantry items to start—build momentum without doing everything on day one
- Quick Start Guide→Get started in minutes with basic meal and receipt tracking
- Your First Meal→Calibration workflow: Weigh your first meal type 3 times to learn your portion sizes (~90 seconds total, then single-tap forever)
- Understanding Pricing→£19.99/month base + £4.99/month per additional member—simple per-person pricing
Meal Logging
3 weigh-ins per meal type, then single-tap forever
- Calibration Workflow (3 Weigh-Ins)→How to calibrate your portion sizes in 1-2 weeks, then never weigh that meal again
- Single-Tap Meal Logging→How 90% of meals become 1-second taps from predicted top 10 list
- Gaussian Portion Models→How 3 weigh-ins create statistical models for fast logging + reconciliation corrects to ground truth
Features
Advanced tools for effortless nutrition tracking
Receipt & Barcode Processing
5-second scan with 2M+ barcodes + Mistral AI
- Receipt OCR Technology→Scan receipts—Mistral AI auto-maps items in 5 seconds across 2M+ preloaded barcodes
- Network Effects: Receipt Database→First user maps item → Saved for all users forever → Month 12: 95%+ cache hit rate
- Intelligent Caching (2M+ Barcodes)→2M+ barcodes preloaded + fuzzy trie matching → Week 1: ~100% manual → Week 8: 100% auto-matched
Reconciliation Engine
Mass conservation physics for research-grade accuracy
- Mass Conservation Physics→How physics achieves ~95% accuracy: consumed = purchased - remaining - waste (ground truth)
- Expiry-Based Auto-Reconciliation→3-tier system: Fast-expiring (auto), slow-expiring (conditional), ultra-stable (rare reweigh)
- Recipe Learning via Reconciliation→How mass conservation learns YOUR recipes: grandmother's sambar (8.5g oil/100ml, not generic 3g)
Privacy & Security
GDPR compliant with EU data hosting
Technical Architecture
Calibration → Single-Tap Logging
- • Week 1-2: Weigh 3 portions per meal type → Gaussian calibration
- • Week 3+: 90% single-tap (1 sec), 9% search (3 sec), 1% weigh (30 sec)
- • Weighted average: 1.5 seconds per meal
- • Weekly effort: 1.6 minutes (vs 42-63 min for competitors)
Data Infrastructure
- • Backend: Java Spring Boot + PostgreSQL
- • Hosting: Railway (EU) for GDPR compliance
- • CDN: Cloudflare global delivery
- • Storage: AWS S3 (EU) with 30-day retention
Mass Conservation Reconciliation
- • Physics equation: Consumed = Purchased - Remaining - Waste
- • Fast-expiring: Milk expires → Auto-reconcile all meals
- • Slow-expiring: Rice consumed? → Prompt reweigh
- • Result: 95% accuracy ground truth (corrects Gaussian estimates)
Intelligent Caching (2M+ Barcodes)
- • 2M+ barcodes preloaded: Instant nutrition lookup for most products
- • Fuzzy trie: OCR-aware edit distance (handles "Teseo" → "Tesco")
- • Mistral AI: First user scans → Maps for all users forever
- • Network effects: Month 12: 95%+ cache hit → 5 sec receipt scan
Can't Find What You're Looking For?
Our support team is here to help with any questions not covered in the documentation.
Ready to Get Started?
Start your free 14-day trial and experience research-grade nutrition tracking.
Start Free Trial