​
Single URL scrape

POST /v1/public/scrape is the main single-page fetch endpoint. Supported format values:
  • markdown
  • raw
  • json
  • metadata
  • screenshot
markdown is the default and is the most agent-friendly option when you want readable content with less markup noise.

​
Batch scrape

POST /v1/public/scrape/batch accepts up to 50 URLs per call.
  • batch requests are always async
  • each URL returns either a queued job entry or an inline validation error
  • callers should poll each returned pollUrl

​
Cache and freshness

URL scraping supports the shared job fields:
  • fresh to bypass cache
  • cacheTtlSec to tune how long the result should be reused
  • webhookUrl for completion callbacks
  • idempotencyKey for safe retries

​
When to choose another endpoint

  • Use /v1/public/scrape/site when you want rendered brand/site signals.
  • Use /v1/public/scrape/company when you want a normalized company profile.
  • Use /v1/public/scrape/rss when the source is already a feed rather than a page.