Connect your agent.
One MCP endpoint gives your agent tools to launch, navigate, inspect, and close browsers.
Set up MCPGive your agents a browser in the cloud. Connect through MCP or the API, automate the web, and watch it happen live.
30 trial minutes · Google sign-in · Alpha access// One connection. A browser ready.
› Open example.com and show me the browser.
One MCP endpoint gives your agent tools to launch, navigate, inspect, and close browsers.
Set up MCPWatch a browser live. Share a temporary signed link or embed the viewer in your own app.
Embed a browserWorkspace-scoped keys, view-only access, expiring sessions, and clear usage allowances.
Understand accessCreate a workspace, connect an agent, and put your trial minutes to work. Everything you need to start is in the docs.
Your application needs a workspace and API key. People opening a signed browser viewer link do not need to sign in. You decide who receives those temporary links.
Yes. Launch a session and connect your agent-browser CLI to its CDP URL. You can also use the built-in MCP tools from a compatible agent client.
Yes. Allow your app’s origin, create a signed link on your backend, and use it in an iframe. Read the embedding guide.
New workspaces start with 30 trial minutes and one browser at a time. You can buy 2,000 prepaid browser minutes for $10 USD plus applicable tax in the dashboard. Sessions have explicit time limits. Persistent profiles, recordings, and subscriptions are not available yet.