Fetch a page
Loads a page in a session and returns its content. Tilion handles rendering, anti-bot recovery, and optional replay capture. Trust content_ok over status: some sites answer with a non-200 while still serving a full document. A fetch that outruns its deadline returns 504; use async_mode and poll /v1/job/{job_id} instead.
Authorizations
Every request carries a bearer API key, which looks like ph_live_.... Keep keys server-side, because a key grants full control of your sessions. A missing header returns 401; an invalid or revoked key currently returns 403, so treat the two together as authentication failure.
Body
An existing session to run the fetch in.
The page to load. Public http/https only.
Execute JavaScript. Set false for static pages; it is faster.
Capture a session replay and return replay_url.
Also return the rendered DOM in html.
Return a job_id immediately instead of waiting. Poll /v1/job/{job_id}.
Response
Successful Response
ok, blocked, error, or pending.
The URL after redirects.
Set only when async_mode was requested.
none, or the challenge family detected.
The page title.
Extracted text content.
Rendered DOM size in bytes.
The rendered DOM. Present only when include_html was set.
A block was detected and recovered automatically.
Attempts, winning strategy, and cost of the recovery.
Present only when record was set.
Challenge classification.
The authoritative signal that a real document came back. Trust this over status.
Set when a blocked status was flipped to ok because real content came back.