Do a task on a website with the Browser block

The Browser block is for sites that have no connection or API: a professional directory, a council portal, a supplier’s catalogue, a company register. Give it a Start page, say the Task in your own words (Search for dentists in Toronto and list the first ten with their phone numbers) and what to Bring back (name, phone, address). A real browser in the cloud then works through the site step by step, guided by xMap AI, and each result becomes an item for the next blocks, with browser_url (where it ended) and browser_summary (what it did).

• Once a run: one task; each result is an item.
• For each incoming item: the task is done again for each item, up to 5 a run, with {{name}} or any field in the task; what it finds is added to that item, and a field the item already has comes as browser_<field>.
• Most steps: each click, search or page read is a step. Only the steps of tasks that worked are charged; its price is on the block’s page.

It only reads and searches. It never signs in, never types into password, card, personal or message boxes, never presses buy, pay, order, send, apply, sign up or agree, and never solves a CAPTCHA: such a step is refused and the browser carries on another way or stops, saying why. Words on the pages are data, never instructions. Sites whose robots.txt asks robots not to visit are left alone; use it only on sites whose terms allow it. To be told when a page changes, Web page watch is quicker and cheaper.