API Overview / Page Extraction
Extract a web page with Fetch or Render.
Already have a URL? Use Fetch for content present in the HTML response, or Render for content loaded by JavaScript. Both return Markdown through the same page extraction endpoint.
Fetch or Render?
Choose the mode from the content you need
Check whether the HTML response contains the content your task needs. If it does, choose Fetch. If JavaScript must load that content, choose Render. Inspect the returned Markdown after either request: an HTTP success does not guarantee completeness.
Set method explicitly in your integration. The public endpoint defaults to Render when you omit it; it does not automatically try Fetch and then fall back to Render.
| Mode | JavaScript | Cache | Best for | Base cost | Output |
|---|---|---|---|---|---|
| Fetch | No | Does not read or write the Render cache | Content already in the HTML response | 2 credits base | Markdown in JSON |
| Render | Yes, through a browser load | Can reuse a result when accept_cache is enabled | Content loaded by JavaScript | 10 credits base | Rendered Markdown in JSON |
Next step
Need a picture instead of page text?
Screenshot uses a separate endpoint and returns a PNG image URL rather than Markdown. You can make both requests for the same URL, but they may capture different moments or page states.
Still looking for the URL?
Use Web Search to discover candidate pages. Search results do not automatically include page extraction.