Browser Automation

Run a Browser Task

Run a natural-language task in a real cloud browser driven by an AI agent — it navigates, clicks, types, scrolls, and extracts ("find the cheapest 65-inch TV on this site and list its specs", "fill the contact form with …"). Submitted in async mode; returns a workflow ID — poll GET /ai/jobs/{workflow_id} (with a delay in between, this can take minutes; the server hard-caps a run at ~330s) for the result. For login-protected tasks pass session_id from List Browser Sessions. Does not execute payments or transfer funds; such tasks are refused.

post/ai/evaluate

Request

  • Base URL: https://api.anakin.io/v1
  • URL: https://api.anakin.io/v1/ai/evaluate
  • Auth: API key in header X-API-Key

Request body

promptstring required

The task in natural language. Be specific about the goal and what to return. Never include passwords or secrets — use session_id for authenticated sites.

asyncboolean

Run asynchronously and return a workflow_id to poll, instead of blocking the request until done (the synchronous mode can block for several minutes). Async is recommended for all clients.

urlstring

Navigate here before starting. Omit to let the agent follow URLs named in the prompt.

session_idstring

Saved browser-session ID (from List Browser Sessions) so the task runs logged in.

max_stepsinteger

Cap on agent steps (navigation/click/type actions).

timeout_msinteger

Task timeout in milliseconds (server caps runs at ~330s regardless).

output_schemaobject

JSON Schema for the result — the agent returns structured data conforming to it.

Response

Task accepted

workflow_idstring

The workflow ID — pass to Get Browser Task Result to poll for completion.

statusstring

Changes

No recorded changes to this operation across all 1 revision of this API.