PRE-ORDER OPEN · LAUNCH Q4 2026

AI food recognition for fitness and nutrition apps.

On-device + iOS.

One photo → dish, portion, kcal, macros.

Built for B2B.

Join the waitlist →

>95%

Top-1 dish accuracy

>468000

Food items USDA

<2 seconds

Response Time

Caesar salad · 320 kcal

Caesar salad

Precision: 98.1%⏱ 0.71с

Ingredients:

· Chicken fillet (fried): 60 g · 100 kcal
· Caesar sauce: 1.5 tbsp · 120 kcal
· Crackers (croutons): 15 g · 60 kcal
· Parmesan cheese: 10 g · 40 kcal
· Lettuce leaves and vegetables: 100 g
· Crackers: 15 tbsp · 40 kcal
· Parmesan cheese: 10 g · 40 kcal
· Lettuce leaves and vegetables: 100 g

320kcal

28gproteins

12gcarbohydrates

18gfats

Add more greens and replace some of the croutons with vegetables.

To lower the glycemic index and keep you feeling fuller longer, double the portion of lettuce (romaine) or add fresh cucumber. This will increase the volume of the dish without adding extra calories and will help your body better digest the saturated fats from the dressing and cheese.

The Best Food Detection NutriVision SDK

The most powerful SDK integration solution for the automatic detection, identification, and monitoring of food products using artificial intelligence.

Hand holding phone over a plated meal — NutriVision detection

7 key benefits of NutriVision SDK:

On-device (GDPR)
Hybrid AI (Vision + LLM)
Portion Estimation (AR/LIDAR)
Data Accuracy (> 95%)
Developer Experience (DX) (10/10)
100% White-Label (Solution)
Offline-First (Architecture)

The NutriVision SDK works quickly (< 2 seconds) and accurately (>95%) to identify products, determine composition, weight, and combine with detailed nutritional information.

Developers using our SDK can create detailed nutritional profiles of users and recommend individual diets, personalized recommendations from a virtual nutritionist, taking into account age, gender, allergens, and goals.

Developer Section (DX)

Scan → Neural Engine → Instant Result. Zero Network. Zero Latency. Zero Cost per Call.

Caesar Salad

Developer-First iOS Integration

Build premium nutrition tracking with zero backend infrastructure. NutriVision On-Device SDK run photos natively via the Apple Neural Engine (ANE), delivering local JSON results in under 85ms.

Real-Time Processing: Fully compatible with Apple AVFoundationcamera streams.
Apple-Native Standards: Outputs normalized coordinates mapped for CoreGraphics CGRect and SwiftUI.
100% Offline Efficiency: Instant macro and ingredient analysis with zero network latency.

Json

{
  "status": "success",
  "session_id": "ses_84f92c11a3b04e",
  "processing_time_ms": 85,
  "media_info": {
    "width": 800,
    "height": 800,
    "format": "JPEG"
  },
  "analysis_results": [
    {
      "dish_name": "Chicken Caesar Salad",
      "confidence_score": 0.984,
      "estimated_weight_g": 250,
      "nutrition_facts": {
        "calories_kcal": 320,
        "macronutrients": {
          "proteins_g": 28.0,
          "carbohydrates_g": 12.0,
          "fats_g": 18.0,
          "dietary_fiber_g": 2.5
        }
      },
      "detected_ingredients": [
        {
          "name": "Grilled Chicken Breast",
          "confidence": 0.99,
          "visual_bounding_box": [0.40, 0.38, 0.76, 0.63]
        },
        {
          "name": "Romaine Lettuce",
          "confidence": 0.96,
          "visual_bounding_box": [0.08, 0.12, 0.96, 0.85]
        },
        {
          "name": "Croutons",
          "confidence": 0.95,
          "count": 7
        },
        {
          "name": "Parmesan Cheese (Shredded)",
          "confidence": 0.89
        },
        {
          "name": "Sun-dried Tomatoes",
          "confidence": 0.94
        },
        {
          "name": "Anchovy Fillets",
          "confidence": 0.87
        },
        {
          "name": "Lime Wedge",
          "confidence": 0.97
        }
      ]
    }
  ]
}