Docs · Extraction

Clean article extraction

Readability extraction — strips nav, ads, sidebars.

POST /v1/extract/clean

Returns the main article body of a page along with the title, word count, and estimated reading time. Built for LLM ingestion and content indexing.

Parameters

Name Type Required Default Description
url string yes — URL of the article.
stealth boolean no true Apply stealth countermeasures.
timeout integer no 30 Per-request timeout.

Request

curl -X POST https://api.datasonar.dev/v1/extract/clean \
  -H "Authorization: Bearer osk_..." \
  -d '{"url": "https://en.wikipedia.org/wiki/Web_scraping"}'

Response

{
  "status": "success",
  "title": "Web scraping - Wikipedia",
  "content_html": "<div>...</div>",
  "content_text": "Method of extracting data from websites...",
  "word_count": 3873,
  "reading_time_min": 16
}

Related

POST /v1/extract/structured
Structured data extraction
POST /v1/scrape
Scrape a single URL