Steel
Three fixed actions — scrape to Markdown/HTML, screenshot, or PDF — one page load in, one response out.
Where Steel lands
Composite scores out of 10, from benchmark runs where every provider faces the identical task corpus. Rank is within that capability's field. Full method: how we measure.
The honest pitch
Reach for it when the job is one URL and the output is page content, a screenshot, or a PDF — no session to keep alive, and no query-string configuration to get right.
Not the pick for a stateful flow: logins, multi-step interaction, a live session to drive, or Steel's own proxy/CAPTCHA handling all sit outside this route.
About Steel
Steel’s Browser Tools are three fixed, single-call actions: /scrape returns a page’s content as HTML, cleaned HTML, Markdown, or Readability-parsed text; /screenshot captures a rendered page, full-page or viewport; /pdf renders one to a PDF. Each is one page load, one response — every control travels in the request body, since this route accepts no query string at all, and the account credential that authenticates each call is injected server-side, never exposed to the caller. Every accepted, forwarded call is billed a flat $0.005, charged the moment the gateway commits to forwarding, on the same NativePort key and balance as the rest of the catalog. Steel’s own Sessions API, CDP browser relay, saved profiles, and proxy/CAPTCHA configuration all sit outside this route — reach for a session-oriented browser product when the job needs more than one page load or state that persists between calls.
Reaching it through NativePort
Identical paths, parameters and responses to Steel's own documentation — the gateway holds the upstream credential and meters your balance at Steel's real published usage price.
curl https://api.nativeport.ai/steel/<native-path> \ -H "Authorization: Bearer $NATIVEPORT_API_KEY"