Docs
Extract

Start a structured data extraction job

POST/extract
AuthorizationBearer <token>

Enter your API key with the bc_ prefix

In: header

Request Body

application/json

TypeScript Definitions

Use the request body type in TypeScript.

agent?

Agent configuration

enableWebSearch?boolean

Enable web search for URL discovery

Defaultfalse
ignoreInvalidURLs?boolean

Ignore invalid URLs and process remaining valid ones

Defaulttrue
ignoreSitemap?boolean

Ignore sitemap during URL discovery

Defaultfalse
includeSubdomains?boolean

Include subdomains in extraction

Defaulttrue
prompt*string

Natural language prompt describing what to extract. Maximum 16384 UTF-8 bytes.

schema?

JSON Schema for structured output. Serialized schema is limited to 65536 UTF-8 bytes.

scrapeOptions?

Scrape options for each URL. zeroDataRetention: true is currently rejected with 400 ZERO_DATA_RETENTION_NOT_SUPPORTED.

showSources?boolean

Include source citations in response

Defaultfalse
urls?array<string>

1–20 unique public HTTP(S) URLs. May be omitted only when enableWebSearch is true. Each URL is limited to 2048 UTF-8 bytes.

Items1 <= items <= 20
webhook?

Webhook configuration

Response Body

application/json

curl -X POST "https://example.com/extract" \  -H "Content-Type: application/json" \  -d '{    "prompt": "Extract all product names, prices, and descriptions from these pages"  }'
{  "id": "string",  "invalidURLs": [    "string"  ],  "success": true,  "url": "string"}
Empty
Empty