Capabilities

Reach any site. Keep every decision.

The platform behind one API: a real browser your agent drives, the accounts you authorize it to use, the guardrails you set — and a compiled skill that replays the next identical run without a model call.

19 capabilities · 4 groups · REST, 35 MCP tools and the dashboard

DELEGATE · ACT · APPROVEYougoal + target URLReal browser runobserve · plan · actControl gateneeds a person?Work donevideo · audit logauthorizeactclearYou decidecode · approval · stopresume in placepauseyou authorize · it does the work · you approve what matters

The platform

19 capabilities, every one shipped

Not a roadmap. Every card below is a mechanism that runs today — open one for the endpoint it exposes, what a call costs, how to switch it on, and the limits it will hit.

The cost engine

Why the second run of a task costs a fraction of the first — and the tenth costs less than that.

Reading the web

Turning a live page into something a model, a pipeline or a person can actually use.

Running and watching

Cadence, observability, and the point where a run stops and asks a person.

Identity, access and the network

The accounts a run may use, the network it comes from, and the pacing that keeps both intact.

Build on the browser layer for agents.

One token-efficient API turns a goal into deterministic, replayable action. Start free and compile your first skill in minutes.