Browser Use Orchestrator icon

Browser Use Orchestrator

Pilote browser-use-cdp-orchestrator (agents web, sessions, jobs, profils)

Actions16

Agent → Fetch Page

AI-generated

Summary

Reads a web page in a remote browser-use browser and returns it as markdown, without an agent loop or LLM. It launches a job on the orchestrator server and, in sync mode, waits for the result.

Inputs

  • url (required) — URL of the page to read, e.g. https://news.ycombinator.com.
  • additionalFields.links — Include link hrefs in the extracted markdown. Defaults to false.
  • additionalFields.profile_id — ID of a browser-use browser profile to load (cookies, localStorage, logins). Create or look up profiles with Create Profile / List Profiles. If omitted, a blank browser is used.
  • additionalFields.proxy_country — Two-letter proxy country code for this job (e.g. fr, us, de), useful for geo-restricted content. If omitted, the server's BROWSER_PROXY_COUNTRY default applies.
  • additionalFields.wait_seconds — Maximum seconds the server waits for the job on the launch call (0–600). 0 launches without waiting. Empty values are not sent, but the 0 default is.
  • executionMode — Synchronous (default) waits for completion, polling Await Job internally. Asynchronous launches the job with wait_seconds forced to 0 and returns a job_id right away.
  • syncMaxWaitSeconds — Sync mode only. Total maximum wait in seconds (default 300). If the job is still running at the deadline, the node returns the current state including job_id, and the job keeps running on the server.

Output shape

One JSON item per input item: the job state/result from the orchestrator (in async mode, a job_id and status).

The exact result fields come from the server and are not defined in the node. Sync mode may return a non-terminal state with job_id on timeout. Terminal statuses are done, error and cancelled. With Continue On Fail enabled, errors are returned as {error: message}. Requires the browserUseOrchestratorApi credential (base URL).

Examples

Example 1: Scrape a page as markdown for a downstream LLM or parser

URL = https://news.ycombinator.com; Additional Fields: Links = true; Execution Mode = Synchronous.

Example 2: Read geo-restricted content from a logged-in session

URL = target page; Additional Fields: Proxy Country = fr, Profile ID = an existing profile id.

Example 3: Fire-and-forget fetch, collect later

Execution Mode = Asynchronous; pass the returned job_id to the Job > Await Job operation.

Links

Discussion