Scraping

POST /scrape takes a JSON body. Only url and proxy_type are required.

Two engines, one request

render_js picks the engine, and the engine decides which other fields mean anything.

• render_js: true (the default) runs a real browser. Use it when the page builds itself client-side, or when you need a screenshot or interaction.
• render_js: false sends a plain HTTP request wearing a real browser’s TLS and HTTP fingerprint. Much cheaper and much faster. Use it whenever the HTML you want is in the first response.

Fields on both engines

Field	Type	Notes
url	string	Required. Validated against SSRF: private and non-HTTP addresses are refused.
proxy_type	datacenter | premium	Required. premium is residential.
render_js	boolean	Default true.
country_code	string	ISO 3166-1 alpha-2. Comma-separate up to 10 (us,gb). See Proxies.
proxy_session	string	Sticky session id, letters and digits, up to 32.
proxy_session_lifetime	string	30s, 30m, 2h. Requires proxy_session.
headers	object	Added to the engine’s own. Yours win on a collision.
cookies	array	Cookie objects sent with the request.
session_id	string	Reuse a stored browser session. 1–128 of [A-Za-z0-9._-].
account_id	string	Use a managed login instead. Mutually exclusive with session_id.

Browser only (render_js: true)

Field	Type	Notes
wait_until	string	Defaults to domcontentloaded if you omit it.
wait	integer	Extra milliseconds to wait, 0–30000.
wait_for	string	A selector to wait for.
viewport_width	integer
viewport_height	integer
block_resources	boolean	Skip images and fonts. Faster and cheaper on bandwidth.
screenshot	full | visible
js_actions	array	Click, fill, scroll, wait, evaluate.

HTTP only (render_js: false)

Field	Type	Notes
method	GET POST PUT PATCH DELETE HEAD	Default GET.
body / json / form	string / any / object	Mutually exclusive. Sending two returns bad_request.
header_order	array	Pins on-the-wire header order — a fingerprinting detail.
follow_redirects	boolean	Default true.
timeout_ms	integer	Caps the whole request.

The response

{
  "ok": true,
  "url": "https://example.com/orders",
  "final_url": "https://example.com/orders",
  "status": 200,
  "headers": { "content-type": "text/html; charset=utf-8" },
  "cookies": [],
  "body": "<!doctype html>...",
  "body_length": 48213,
  "body_encoding": "utf-8",
  "truncated": false,
  "content_type": "text/html; charset=utf-8",
  "elapsed_ms": 1840,
  "blocked": false,
  "block_reason": null,
  "redirects": []
}

Two fields are easy to miss:

• body_encoding is utf-8 or base64. Binary responses, and text we could not decode, come back base64. Check it before parsing.
• truncated means the body hit the size cap and was cut. body_length is the byte length after truncation.

When ok is false, the shape is different — see Errors.