Browser agents are automation systems that read pages, click controls, fill forms, upload files, and navigate workflows through a browser runtime.
They matter when the real system of record lives behind web UI, authentication, anti-bot checks, or approval steps that an API does not expose.
Why it matters
- Browser agents can interact with the same workflows humans use.
- They need stronger state, evidence, and recovery controls than simple API clients.
- In Steel, browser agents usually run on top of Sessions, Profiles, Credentials, and Files.