Anakin icon

Anakin

Scrape websites, search with AI, and extract structured data

Actions108

Browser → Browser Task

AI-generated

Summary

Run a natural-language browser automation task using a real AI-driven browser, specifying the goal in plain text to navigate and interact with websites and extract information as instructed.

Inputs

  • Prompt (required) — The natural language task description that specifies what the AI browser should accomplish. Be specific about the goal and what output to return. Do not include passwords or secrets; instead use Session ID for authenticated sites.
  • Additional Options — Optional settings to control task behavior, including URL to navigate to before starting, session ID for authenticated sessions, max number of navigation/click/type steps, max wait time for task completion, poll interval for status checks, output JSON schema to define structured result format, and timeout in milliseconds for the task.

Output shape

a single JSON object representing the completed browser automation task result

The node submits the task asynchronously and polls the service until completion, failure, timeout, or error occurs. On success, it returns structured data that may include fields matching the provided output schema if specified. On failure or timeout, an error is thrown or returned as failure in the output if 'Continue on Fail' is enabled.

Examples

Example 1: Find the cheapest 65-inch TV on a shopping site

Provide a detailed prompt like 'Find the cheapest 65-inch TV on this site and list its specs', optionally specify the site URL, and configure max wait time or polling intervals as needed.

Links

Discussion