⚙️

Image Classification

ActiveCoreMUSTFREE

What it does

Classify products directly from photos using AI vision analysis. Upload an image URL or base64-encoded image data, and POTAL will identify the product and assign the correct HS code. Supports JPEG, PNG, GIF, and WebP formats up to 5MB. Ideal for visual product catalogs or mobile-first workflows.

Required Fields for 100% Accuracy

0 required / 4 total

Field	Type	Required	Description	Example
`imageUrl`	string	opt	URL of the product image (JPEG, PNG, GIF, WebP)Use a clear, well-lit photo with the product centered. Provide either imageUrl or imageBase64.	`https://example.com/product.jpg`
`imageBase64`	string	opt	Base64-encoded image dataMax 5MB after decoding. Use for direct uploads without a hosted URL.	`/9j/4AAQSkZ...`
`productHint`	string	opt	Optional hint to guide AI classificationImproves accuracy for ambiguous images — e.g., "handbag" vs "luggage".	`kitchen appliance`
`destination_country`	string	opt	Destination country for HS10 selection	`US`

Tips for Best Accuracy

✓Either imageUrl or imageBase64 must be provided (one is required)
✓Clear, single-product photos yield the best results
✓productHint significantly improves accuracy for visually ambiguous products
✓Image classification uses AI vision — slightly slower than text-based classification

Common Mistakes

✗Providing both imageUrl and imageBase64 — use one or the other
✗Low-quality or cluttered images — crop to show only the product
✗Images with text overlays or watermarks — these confuse the vision model
✗Expecting the same confidence as text-based — image mode typically scores 0.7–0.9

How to use it

Prepare your image

Use a clear product photo. Supported formats: JPEG, PNG, GIF, WebP. Max size: 5MB.

Send as URL or base64

Include imageUrl for hosted images or imageBase64 for direct upload in the classify request.

Review detected product

The response includes detectedProductName showing what the AI identified from the image.

Refine if needed

Add a productHint to guide classification if the image is ambiguous.

API Reference

POST/api/v1/classify

Request

{
  "imageUrl": "https://example.com/product-photo.jpg",
  "productHint": "kitchen appliance"
}

Response

{
  "hsCode": "8516.79",
  "description": "Electric coffee maker",
  "confidence": 0.87,
  "detectedProductName": "Electric drip coffee maker",
  "alternatives": [
    {
      "hsCode": "8419.81",
      "confidence": 0.62
    }
  ]
}

cURL

curl -X POST https://potal.app/api/v1/classify \
  -H "X-API-Key: pk_live_your_key" \
  -H "Content-Type: application/json" \
  -d '{"imageUrl":"https://example.com/product-photo.jpg","productHint":"kitchen appliance"}'

Related Features

⚙️HS Code Classification

Classify products into 5,000+ HS codes using the v3.3 GRI pipeline

⚙️Confidence Score

Multi-dimensional confidence scoring for every classification result

⚙️Batch Classification

Classify hundreds of products at once via CSV or JSON batch API

⚙️Origin Detection

Detect country of origin from product description and brand analysis

⚙️HS Code Validation

9-field input validation ensuring accurate classification inputs

Having issues?

🐛 Report a Bug ❓ Ask a Question 💡 Suggest Improvement