Web page to Markdown API (WEB_SCRAPE)

WEB_SCRAPE reads one page and returns its main content as Markdown: headings, paragraphs, lists, tables, links and images, with the site’s header, footer and navigation removed. Pages that script draws, such as shop product pages, are rendered before they are read. To read a search result, pass a link from GOOGLE_SERP’s organicResults.


`taskType`	`WEB_SCRAPE`
Input	`url` (required); `maxChars`, `country`
Credits	1 per task, a multi-page PDF included
Country	`US`, or `KR` for Coupang and Olive Young storefronts, when you send none.
Location	Not supported (`422 LOCATION_UNSUPPORTED`).

Request

curl -X POST https://api.querying.ai/v1/async/task \
  -H "Authorization: Bearer $QUERYING_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "taskType": "WEB_SCRAPE", "payload": { "url": "https://www.example.com/guide", "maxChars": 50000 } }'

The task runs asynchronously. Read the result as described in Async tasks.

Parameters

Field	Required	Description
`url`	Yes	A public `http` or `https` address on the default port, up to 2,048 characters. Addresses with credentials, private or internal hosts, or other ports return `422 VALIDATION_ERROR`.
`maxChars`	No	Caps `markdown`, an integer from 1,000 to 200,000. Defaults to 100,000.
`country`	No	The language and region the page is requested for. Without it, domestic Coupang and Olive Young storefronts use `KR` and every other page uses `US`.

Response

For example:

{
  "url": "https://www.example.com/guide",
  "finalUrl": "https://www.example.com/guide",
  "statusCode": 200,
  "title": "...",
  "description": "...",
  "author": "...",
  "publishedAt": "2026-09-12T07:36:36Z",
  "siteName": "...",
  "language": "en",
  "image": "https://...",
  "markdown": "# ...\n\n...\n\n![...](https://...)",
  "images": [{ "url": "https://...", "alt": "..." }],
  "truncated": false
}

Field	Description
`url`	The address you sent.
`finalUrl`	The address after redirects.
`statusCode`	The HTTP status of the page that was read.
`markdown`	The main content. Never empty in a completed task.
`images`	The pictures in `markdown`, in the order they appear.
`truncated`	`true` when `markdown` stopped at `maxChars`. The cut falls on a paragraph break where possible.
`title`, `description`, `author`, `publishedAt`, `siteName`, `language`, `image`	Present when the page declares them.

PDFs

HTML, plain text and text-based PDFs are supported. A PDF returns its text in page order, its title, author, subject (as description) and language when present, and images: []. OCR, attachments and layout reconstruction are not supported.

PDF limit	Value
Download size	8 MiB
Pages	100
Extracted text	1,000,000 characters
Extraction time	20 seconds

The whole document is checked regardless of maxChars; only the returned Markdown is shortened. Textless, encrypted, damaged, incomplete and over-limit PDFs fail with PAGE_UNAVAILABLE.

Limits and errors

A page that cannot be read fails the task with PAGE_UNAVAILABLE and the reason: an unsupported format, a page answering 404 or 410, or a site that refuses the request.
A redirect to an address that breaks the url rules fails the task. The rules are checked again on every redirect hop.
WEB_SCRAPE ignores markdown, rawResponse and searchQueries. It rejects html and analysis with 422 and citationAnalysis with 400; see unsupported flags.
Other request errors are listed in Errors.

Web crawl to read many pages of one site
Web map to list a site’s addresses first

Get started

Core concepts

Answer engines

Search and web data

Request options

Research

GEO monitoring

Account and data

Web page to Markdown API (WEB_SCRAPE)

Request

Parameters

Response

PDFs

Limits and errors

​Request

​Parameters

​Response

​PDFs

​Limits and errors

​Related

Request

Parameters

Response

PDFs

Limits and errors

Related