Image Classification

This page shows how to use the VehiCore Image Classification API to detect what’s in an image (e.g. car vs document) using AI.

You can send an image in two ways:

• Upload: multipart/form-data with field image
• URL: pass ?url=... (GET, or POST without upload)

Authentication is required via X-API-Key. To get a key, register and purchase at vehicore.pro/api-keys.

How it works: You send an image (file upload or ?url=...) and your X-API-Key. The service runs a classification model (CLIP or ViT) and returns the best-matching label plus scores.

Classify image (CLIP — zero-shot labels)

POST /api/image-classification/clip

CLIP runs zero-shot classification against a list of labels (default labels are: car, document, motorcycle, truck, person, other).

Request

Headers

Body (multipart/form-data)

Query parameters

Example request

curl -X POST "https://api.vehicore.pro/api/image-classification/clip?labels=car,document,motorcycle,truck,person,other" \
  -H "X-API-Key: YOUR_API_KEY" \
  -F "image=@/path/to/image.jpg"

import fs from 'node:fs';

const formData = new FormData();
const buf = fs.readFileSync('/path/to/image.jpg');
formData.append('image', new Blob([buf], { type: 'image/jpeg' }), 'image.jpg');

const url = new URL('https://api.vehicore.pro/api/image-classification/clip');
url.searchParams.set('labels', 'car,document,motorcycle,truck,person,other');

const response = await fetch(url, {
  method: 'POST',
  headers: { 'X-API-Key': process.env.VEHICORE_API_KEY },
  body: formData,
});

const data = await response.json();

import os
import requests

url = "https://api.vehicore.pro/api/image-classification/clip"
headers = {"X-API-Key": os.environ["VEHICORE_API_KEY"]}
files = {"image": open("/path/to/image.jpg", "rb")}
params = {"labels": "car,document,motorcycle,truck,person,other"}

response = requests.post(url, headers=headers, files=files, params=params)
print(response.json())

Response

On success the API returns:

{
  "result": "car",
  "confidence": 0.93,
  "model": "clip",
  "predictions": [
    { "label": "car", "score": 0.93 },
    { "label": "document", "score": 0.03 }
  ]
}

See Swagger for the full schema.

Possible errors

All error responses use the shape { "message": "Human-readable text", "code": "OPTIONAL_CODE" }. See Error responses for details.

Classify image (ViT — ImageNet)

POST /api/image-classification/vit

ViT runs ImageNet classification (fixed label space). VehiCore maps common vehicle/document-like outputs into its categories when possible.

Request

Same as CLIP, except labels is ignored.

Headers

Body (multipart/form-data)

Query parameters

Example request

curl -X POST "https://api.vehicore.pro/api/image-classification/vit" \
  -H "X-API-Key: YOUR_API_KEY" \
  -F "image=@/path/to/image.jpg"

URL mode (GET)

You can also call the API using GET with a URL:

• GET /api/image-classification/clip?url=...&labels=...
• GET /api/image-classification/vit?url=...

Example:

curl -H "X-API-Key: YOUR_API_KEY" \
  "https://api.vehicore.pro/api/image-classification/clip?url=https://example.com/image.jpg&labels=car,document,other"