FoodLens · CaloAI REST API

FoodLens API Docs

The FoodLens (CaloAI) REST API is a food image recognition API. Send one food photo, and for every food in it you get back the name, position, serving size, ingredients, calories, and 36 nutrients as JSON. This page introduces the features and the response structure.

Overview

CaloAI is the FoodLens 2.0 engine. Its response format is the same as FoodLens, and some fields (such as food groups) are filled in only by CaloAI.

Features and endpoints

Paths are relative, without the server address.

Android · iOS SDK

To build FoodLens into an app, you can use the SDKs instead of calling the REST API. The UI SDK includes ready-made camera, result, and food search screens; the Core SDK lets you build your own screens. SDK results use the same response format as this document.

Common response format

Every response is application/json and wrapped in the structure below.

Recognition response

body contains one food object for each food in the photo. The example below is the result for one plate of prawn oil pasta (values are examples).

Food object

Candidate object

Missing values are returned as -1 for numbers and as an empty string for strings.

Example response

{
  "body": [
    {
      "name": "Oil Pasta",
      "fullName": "Prawn Oil Pasta",
      "position": { "xmin": 27, "xmax": 712, "ymin": 43, "ymax": 508 },
      "candidates": [
        {
          "id": -1,
          "manufacturer": "",
          "foodName": "Oil Pasta",
          "fullFoodName": "Prawn Oil Pasta",
          "foodType": "",
          "servingSize": 388.0,
          "totalServingSize": 388.0,
          "unit": "g",
          "counts": "plate(L)",
          "countsDetail": [
            { "countType": "serving", "servingSize": 388.0, "countsValue": 1.0, "counts": "plate(L)", "subCountsValue": -1.0, "subCounts": "" }
          ],
          "ingredients": [
            { "name": "Noodles", "gram": 250.0 },
            { "name": "Prawn", "gram": 80.0 },
            { "name": "Oil", "gram": 50.0 },
            { "name": "Parmesan", "gram": 5.0 },
            { "name": "Salt", "gram": 2.5 },
            { "name": "Black pepper", "gram": 0.5 }
          ],
          "energy": 872.21,
          "carbohydrate": 68.36,
          "protein": 25.46,
          "fat": 51.9,
          "totalSugars": 0.0,
          "totalDietaryFiber": 0.0,
          "calcium": 117.31,
          "iron": 2.96,
          "magnesium": 2.55,
          "phosphorus": 220.11,
          "potassium": 363.71,
          "sodium": 1043.66,
          "zinc": 0.21,
          "selenium": 1.78,
          "retinol": 12.75,
          "betaCarotene": 5.54,
          "vitaminA": 13.21,
          "thiamin": 0.15,
          "riboflavin": 0.1,
          "niacin": 4.27,
          "vitaminB6": 0.0,
          "biotin": 0.05,
          "totalFolate": 0.38,
          "vitaminC": 0.8,
          "vitaminD": 0.03,
          "vitaminE": 40.68,
          "vitaminK": 86.14,
          "isoleucine": 74.42,
          "leucine": 141.49,
          "valine": 95.69,
          "bcaa": 5.21,
          "cholesterol": 5.3,
          "saturatedFattyAcid": 6.26,
          "epa": 0.0,
          "dha": 0.0,
          "omega3FattyAcid": 3.41,
          "omega6FattyAcid": 17.07,
          "transFattyAcid": 0.04,
          "grainGroup": -1.0,
          "proteinGroup": -1.0,
          "vegetableGroup": -1.0,
          "fruitGroup": -1.0,
          "oilGroup": -1.0,
          "sugarGroup": -1.0,
          "dairyGroup": -1.0
        }
      ],
      "ingredients": [
        { "name": "Noodles", "gram": 250.0 },
        { "name": "Prawn", "gram": 80.0 },
        { "name": "Oil", "gram": 50.0 },
        { "name": "Parmesan", "gram": 5.0 },
        { "name": "Salt", "gram": 2.5 },
        { "name": "Black pepper", "gram": 0.5 }
      ]
    }
  ],
  "path": "",
  "success": true,
  "code": ""
}

Nutrient fields

All nutrient values are per one serving (servingSize). Missing values are returned as -1.

Food info response

Looking up a food by ID returns that one food in the same format as the recognition response. Since there is no photo, every position value is 0 and ingredients is an empty list.

{
  "body": [
    {
      "name": "Oil Pasta",
      "fullName": "Prawn Oil Pasta",
      "position": { "xmin": 0, "xmax": 0, "ymin": 0, "ymax": 0 },
      "candidates": [
        {
          "id": -1,
          "manufacturer": "",
          "foodName": "Oil Pasta",
          "fullFoodName": "Prawn Oil Pasta",
          "foodType": "",
          "servingSize": 388.0,
          "totalServingSize": 0.0,
          "unit": "g",
          "counts": "plate(L)",
          "countsDetail": [
            { "countType": "serving", "servingSize": 388.0, "countsValue": 1.0, "counts": "plate(L)", "subCountsValue": -1.0, "subCounts": "" }
          ],
          "energy": 872.21,
          "carbohydrate": 68.36,
          "protein": 25.46,
          "fat": 51.9,
          "totalSugars": 0.0,
          "totalDietaryFiber": 0.0,
          "sodium": 1043.66,
          "cholesterol": 5.3,
          "saturatedFattyAcid": 6.26,
          "transFattyAcid": 0.04,
          "grainGroup": -1.0,
          "proteinGroup": -1.0,
          "vegetableGroup": -1.0,
          "fruitGroup": -1.0,
          "oilGroup": -1.0,
          "sugarGroup": -1.0,
          "dairyGroup": -1.0
        }
      ],
      "ingredients": []
    }
  ],
  "path": "",
  "success": true,
  "code": ""
}

The example shows only some nutrients to save space. Actual responses include every field listed in Nutrient fields.

Food search response

The food names, IDs, and values below are examples that show the format.

Simple list

Returns only the food ID, name, and manufacturer. Get the details of the chosen food with Food info.

{
  "body": [
    { "id": 6699, "foodName": "Steamed Steak", "manufacturer": "" },
    { "id": 6702, "foodName": "Grilled Sirloin Steak", "manufacturer": "" },
    { "id": 789, "foodName": "Pork Chop Steak", "manufacturer": "" }
  ],
  "success": true,
  "code": ""
}

Detailed list

Also returns the serving size, calories, and serving unit. Numeric values are returned as strings. Names and units come in the result language; the example below shows Korean results.

{
  "body": [
    { "id": 360021, "foodName": "연어", "manufacturer": "부경수산", "weight": "100g", "energy": "261.0", "countsValue": "1", "counts": "총 내용량", "subCountsValue": "", "subCounts": "" },
    { "id": 94911, "foodName": "연어초밥", "manufacturer": "", "weight": "180g", "energy": "287.5", "countsValue": "1", "counts": "인분", "subCountsValue": "6", "subCounts": "개" },
    { "id": 94912, "foodName": "연어감자조림", "manufacturer": "", "weight": "125g", "energy": "140.24", "countsValue": "1", "counts": "접시(중)", "subCountsValue": "", "subCounts": "" }
  ],
  "success": true,
  "code": ""
}

Error responses

Calculating values

• Ingredient weights are also per one serving.
• Value per 100 g = nutrient value × 100 ÷ servingSize
• Values for the whole package: when countsDetail has an item with countType: "total", convert using that item's servingSize.
• A value of -1 means no data, so do not treat it as 0.

Getting started

1. Contact us at sales@doinglab.com or +82-2-6959-5657.
2. After the contract, you receive the full specification with your API key, server address, and request parameters.
3. Develop and validate during the 3-month free trial (up to 3,000 FoodLens calls).
4. To add it to your app quickly, use the Android and iOS SDKs with Sample UI. Manuals: Android · iOS

See FoodLens pricing. To try recognition first without signing up, upload a food photo in the Diet Camera app (available in Korea).