We're releasing Browser Agent, a new capability that lets your AI agents interact with any website.

What it does

Browser Agent can navigate the web just like a human would. Fill forms, click buttons, extract data, and handle multi-step flows across any site.

This unlocks automation for the long tail of apps that don't have APIs. Connect your workflows to legacy systems, niche tools, or any web interface.

How it works

Under the hood, Browser Agent uses a headless browser controlled by our AI. It sees the page, understands the context, and takes actions based on your instructions.

const result = await browserAgent.run({
  task: "Log into Salesforce and export all leads from Q4",
  credentials: { username, password }
});

The agent handles CAPTCHAs, popups, and dynamic content automatically.

Getting started

Browser Agent is available today for all teams on the Pro plan.