Working with Captures - Extralt Docs

A capture is a single extracted data record. Each run produces one capture per successfully extracted URL.

Fetching captures

List captures from a specific run:

curl -s "https://api.extralt.com/captures?runId=$RUN_ID" \
  -H "Authorization: Bearer $EXTRALT_API_KEY" | jq

Capture structure

Each capture contains:

Field	Description
`_id`	Unique capture identifier
`url`	The source URL that was extracted
`data`	Extracted product data (base fields + optional details)
`robotId`	The robot that extracted it
`runId`	The run that produced it

Example capture

{
  "_id": "cap_abc123",
  "url": "https://example-store.com/products/sneakers",
  "data": {
    "title": "Air Max 90",
    "brand": "Nike",
    "description": "Classic sneaker with visible Air cushioning.",
    "images": [
      "https://example-store.com/images/am90-1.jpg",
      "https://example-store.com/images/am90-2.jpg"
    ],
    "identifiers": {
      "gtin": "1234567890123",
      "mpn": "DN3707-100"
    },
    "variants": [
      {
        "option_values": ["10", "White"],
        "sku": "AM90-WHT-10",
        "price": { "amount": 130, "currency": "USD" },
        "availability": "in_stock"
      },
      {
        "option_values": ["11", "White"],
        "sku": "AM90-WHT-11",
        "price": { "amount": 130, "currency": "USD" },
        "availability": "out_of_stock"
      }
    ],
  }
}

Filtering

You can filter captures by run:

GET /captures?runId=your-run-id

Pagination

When a run produces many captures, results are paginated. See Common Patterns for pagination handling.

Bulk download

For large datasets, use the run download endpoint instead of paginating through individual captures. It returns a compressed .jsonl.lz4 file with all captures from the run.

See Running Extractions for download details.

What's next

Running Extractions -- bulk download via signed URL
Common Patterns -- pagination and error handling
API Reference -- full captures endpoints