Every layer is proprietary. The intelligence core cannot be assembled from off-the-shelf APIs. It requires training on order behavior and biometric correlation.
Data in. One meal recommendation out. Every column is a distinct proprietary system. The intelligence layer in the center is why this cannot be replicated from APIs alone.
Apple Health · Whoop · Oura · Garmin · Samsung
Steps · Calories · HR · HRV · Sleep · Recovery · Strain · VO2max
Age · Gender · Height · Weight · Activity level · Goals
Cuisine preferences · Allergies · Restrictions · Meal frequency · Budget range
365 days of activity, sleep, calories, and macros granted on signup
Past Aevia orders + imported history from Uber Eats, DoorDash, Instacart where available
Dexcom / Libre · Real-time glucose · Meal spike patterns
Real-time sync via Terra API + HealthKit · Updates every 15 min · HRV and recovery recalculate targets 2x daily
365-day archive processed into avg daily cals, macro ratios, time-of-day patterns, activity norms, BMR, and TDEE
Cuisine, dietary restrictions, meal frequency, budget, and ordering patterns consolidated into scoring weights
Proprietary dish catalog with verified macro and micronutrient data. Every dish pre-indexed before the app opens. 500+ restaurants catalogued at launch. AI vision + public databases + manual verification.
Build target → Measure gap → Score dishes
Macro, micro, caloric targets built from biometrics, demographics, and historical patterns. Adjusted by hour of day.
Real-time target vs. consumed. Surplus and deficit correction in both directions.
Menu match scoring (0–100 per dish). Recovery window detection and nutrient timing.
Partner availability matching. Preference weighting from order history.
Uber Eats · DoorDash · Any local restaurant in range · Real-time menu + ETA feed · In-restaurant ordering via QR or direct
ACKitchen · Sakara Life · Thistle · Factor · Trifecta · Local premium prep partners
Instacart · Amazon Fresh · Whole Foods · Phase 2
Sweetgreen · Chipotle · Real-time via Nutritionix · 500+ chains indexed
Employer wellness · Clinics · White-label Phase 3
A personalized meal planning system that builds a target nutrition profile per user, per meal window. The target recalculates multiple times per day. Every confirmed order trains the preference model.
Flywheel: Each confirmed order feeds back into the preference model. Which recommendations the user accepts, which they skip, what time they order, which cuisines they gravitate toward. Scoring accuracy improves with every interaction.
A proprietary catalog of verified macro, micro, and calorie data for every dish across every fulfillment channel. Pre-indexed before the app opens. Grows with every market and every restaurant partner added.
No single data provider covers all five fulfillment channels at dish level. Aevia does. The database compounds with every market launch and every restaurant added, building a verified nutritional layer that is not available anywhere else and cannot be replicated from a single source.
The market has nutrition trackers, meal planners, biometric platforms, and delivery apps. Each covers one or two of these capabilities. None covers all five.
| Capability | AEVIA | Levels | MyFitnessPal | Noom | DoorDash / Uber Eats |
|---|---|---|---|---|---|
| Real-time biometric ingestion Wearables, CGM, blood panels synced continuously | ✓ | ✓ | Activity only | — | — |
| Dynamic per-meal nutrition target Recalculates by meal window based on activity, recovery, intake | ✓ | Daily only | Static daily | Static daily | — |
| Verified dish-level nutrition database Pre-indexed macros/micros for restaurant dishes, not user-submitted | ✓ | — | User-submitted | — | Optional calorie ranges |
| Gap-to-order signal Scores every available dish against current body state, outputs one recommendation | ✓ | — | — | — | — |
| Multi-channel ordering Delivery, pickup, dine-in, meal prep, grocery in one app | ✓ | — | — | — | Delivery + pickup only |
Track what you already ate. 200M+ users across category. Large food databases (14M+ items in MFP, mostly user-submitted). Connect to wearables for activity data. No connection to food supply chain. No forward-looking recommendations. Look backward, not forward.
Generate weekly meal plans with grocery lists. Some integrate Instacart or Amazon Fresh for grocery delivery. None integrate with restaurant delivery platforms. No biometric ingestion. No real-time adaptation. Plan meals for home cooking, disconnected from body state and the restaurant supply chain.
Levels ($67M raised, a16z) is closest: CGM + blood panels + AI food logging + meal scoring. But no dish-level database, no menu indexing, no ordering. Noom ($540M raised) is behavioral coaching with food logging. Season Health ($45M raised, a16z) is a clinical "food pharmacy" with dietitian-led plans and delivery. Insurance-reimbursed, not real-time, not consumer-facing at scale.
$430B global GOV. Massive order volume, zero nutritional intelligence. Only 27% of DoorDash menus and 19% of Uber Eats menus display calorie counts (Tufts/CSPI, 2023). FDA has not enforced nutrition labeling on third-party platforms. Restaurants can optionally add calorie ranges via API. No scoring, no targeting, no health context.
The moat is not any single capability. It is the closed loop: biometric data in, nutrition target calculated, every dish scored, order routed. Built on a proprietary database that no competitor has and no single API provides.
Biometric ingestion + nutrition targeting + dish database + gap scoring + multi-channel ordering. No competitor has all five.