Classification

Classify text and image datasets using custom labels.

POST /v1/classification

Javascript

import { JigsawStack } from "jigsawstack";

const jigsaw = JigsawStack({ apiKey: "your-api-key" });

const response = await jigsaw.classification({
  "dataset": [
        {
              "type": "image",
              "value": "https://jigsawstack.com/preview/classification-example-1.jpg"
        },
        {
              "type": "image",
              "value": "https://jigsawstack.com/preview/classification-example-2.jpg"
        }
  ],
  "labels": [
        {
              "type": "text",
              "value": "hotdog"
        },
        {
              "type": "text",
              "value": "not a hotdog"
        }
  ]
})

Response

{
  "success": true,
  "predictions": [
        "not a hotdog",
        "hotdog"
  ],
  "_usage": {
        "input_tokens": 69,
        "output_tokens": 14,
        "inference_time_tokens": 3942,
        "total_tokens": 4025
  }
}

Body

dataset · array · required

Array of data samples to classify. Each sample contains:

• type: Either “text” or “image”
• value: The content to classify (text string or image URL)

Maximum of 32 samples allowed per request. At least 1 sample required.

All items in the dataset must be of the same type. You cannot mix text and image samples in a single request.

labels · array · required

Array of classification labels. Each label contains:

• key: Optional identifier for the label (string or number). Labels sharing the same key are merged into a single combined label, so you can attach multiple text descriptions and/or example images to the same class.
• type: Either “text” or “image”
• value: The label content (text string or image URL)
• text: Optional text description for an image label, used to give the model extra context about what the image represents.

Maximum of 24 labels allowed per request. At least 2 distinct labels required (labels merged by key count as one).

multiple_labels · boolean · default: "false"

Whether to allow multiple labels per classification result. When false, each prediction returns a single label. When true, each prediction can return multiple labels.

task · string

Optional natural-language instruction that describes the classification task (e.g. “Classify the sentiment of the following reviews”). When provided, it is prepended to the prompt to steer the model.

Header

x-api-key · string · required

Your JigsawStack API key

Response

success · boolean

Indicates whether the call was successful.

_usage · object

Usage information for the API call.

Show _usage

input_tokens · number

Number of input tokens processed.

output_tokens · number

Number of output tokens generated.

inference_time_tokens · number

Number of tokens processed during inference time.

total_tokens · number

Total number of tokens used (input + output).

log_id · string

A unique identifier for the request

predictions · array[string | string[]]

Array of classification results. Each element corresponds to a dataset sample:

• If multiple_labels is false: Each element is a string (single label)
• If multiple_labels is true: Each element is an array of strings (multiple labels)

Javascript

import { JigsawStack } from "jigsawstack";

const jigsaw = JigsawStack({ apiKey: "your-api-key" });

const response = await jigsaw.classification({
  "dataset": [
        {
              "type": "image",
              "value": "https://jigsawstack.com/preview/classification-example-1.jpg"
        },
        {
              "type": "image",
              "value": "https://jigsawstack.com/preview/classification-example-2.jpg"
        }
  ],
  "labels": [
        {
              "type": "text",
              "value": "hotdog"
        },
        {
              "type": "text",
              "value": "not a hotdog"
        }
  ]
})

Response

{
  "success": true,
  "predictions": [
        "not a hotdog",
        "hotdog"
  ],
  "_usage": {
        "input_tokens": 69,
        "output_tokens": 14,
        "inference_time_tokens": 3942,
        "total_tokens": 4025
  }
}