trackup-food-analyze（追踪食物分析）

TrackUp Food Analyze

Use this skill to analyze a food image with:

• https://deepeat.ai/step.aispark.api.API/AnalyzeWholeFood
• https://deepeat.ai/step.aispark.api.API/ExtractFoodMacros
• https://deepeat.ai/step.aispark.api.API/AnalyzeIngredients
• https://deepeat.ai/step.aispark.api.API/GetHealthInsight
• https://deepeat.ai/step.aispark.api.API/SearchFood

Rules

• Always use POST
• Always send Content-Type: application/json
• Image analysis requests must contain image_url or image_base64
• If both exist, image_base64 wins
• image_base64 must be raw base64 without data:image/...;base64,
• For complete food image analysis, call AnalyzeWholeFood directly
• Only call a specialized image-analysis API when the user explicitly asks for one specific analysis dimension
• Use ExtractFoodMacros for basic macro extraction
• Use AnalyzeIngredients for deep ingredient and metabolic analysis
• Use GetHealthInsight for professional health insight
• SearchFood request body must contain keyword
• SearchFood returns at most the first 20 results

API Selection

Choose the API based on the user's intent:

1. Full food analysis: AnalyzeWholeFood
2. Basic macro extraction only: ExtractFoodMacros
3. Deep ingredient and metabolic analysis only: AnalyzeIngredients
4. Professional health insight only: GetHealthInsight
5. Keyword-based lookup without image analysis: SearchFood

Food Search

Use SearchFood when the user wants keyword-based food lookup instead of image analysis.

Request example:

curl --silent --show-error --location \
  'https://deepeat.ai/step.aispark.api.API/SearchFood' \
  --header 'Content-Type: application/json' \
  --data '{
    "keyword": "banana"
  }'

Expected response shape:

{
  "foods": [
    {
      "food_name": "Banana",
      "calories_cal": "89000",
      "weight_g": "100",
      "carbs_g": "22.8",
      "proteins_g": "1.1",
      "fats_g": "0.3",
      "display_gi": 51,
      "brand": ""
    }
  ],
  "pagination": {
    "size": 20,
    "page": 1,
    "total": 123
  }
}

Local Image Workflow

Convert local image to base64:

base64 < /absolute/path/to/image.jpg | tr -d '\
' > /tmp/food_image.b64

Build payload:

printf '{"image_base64":"' > /tmp/food_image.json
cat /tmp/food_image.b64 >> /tmp/food_image.json
printf '"}' >> /tmp/food_image.json

Call full food analysis directly:

curl --silent --show-error --location \
  'https://deepeat.ai/step.aispark.api.API/AnalyzeWholeFood' \
  --header 'Content-Type: application/json' \
  --data-binary @/tmp/food_image.json

Proxy Fix

If curl incorrectly tries a local proxy like 127.0.0.1:7890, retry with:

env -u http_proxy -u https_proxy -u HTTP_PROXY -u HTTPS_PROXY -u ALL_PROXY curl ...

Response Fields

AnalyzeWholeFood:

• food_name
• total_weight_g
• total_calories_kcal
• carbs_g
• proteins_g
• fats_g
• display_gi
• dietary_fiber_g
• sugar_g
• ingredients[]
• health_score
• health_profile

ExtractFoodMacros:

• food_name
• total_weight_g
• total_calories_kcal
• carbs_g
• proteins_g
• fats_g

AnalyzeIngredients:

• display_gi
• dietary_fiber_g
• sugar_g
• ingredients[]

GetHealthInsight:

• health_score
• health_profile

SearchFood:

• foods[]
• pagination.page
• pagination.size
• pagination.total

health_score meaning:

• 1-5 score generated by AI food evaluation
• 5 is the highest score
• Use it to help users make nutritionally informed choices

Output Rules

• Keep numeric values exactly as returned
• Do not invent missing fields
• For complete image analysis, use AnalyzeWholeFood as the default API
• For specialized analysis, use only the dedicated API that matches the user's request
• For SearchFood, return the foods in API order and do not claim more than 20 items unless the response actually contains them

Common Errors

• 400 invalid_request
• 422 food_not_recognized
• 422 barcode_only_result
• 429 rate_limit_exceeded
• 500 internal_error
• 502 upstream_flow_error