Skip to content
Browserberg Browserberg

Browserberg documentation

Documentation for EU-sovereign browser infrastructure for AI agents: sessions, verbs, workflows, the credential vault, audit and the benchmark.

Stylised overview of the Browserberg platform: browser sessions in hardened containers, driven through one API

What Browserberg is

Browserberg is EU-sovereign browser infrastructure for AI agents. Your agent gets a real Chromium in a hardened container and drives it through three verbs — observe, act, extract — or hands over whole tasks to a hosted agent loop. The platform is authenticated-first: built for your own accounts and partner portals, with no stealth, no fingerprint spoofing and no CAPTCHA solving, ever.

These docs cover the whole surface: getting a first session running, the concepts behind sessions, verbs and provenance, hands-on guides for workflows, credentials and triggers, the full API reference, and the open benchmark we measure ourselves against.

New here? Start with the quickstart under Getting started — an API key and a few minutes take you from nothing to extracted data. If you already know what you are after, the reference section is organised by endpoint.

Find your way in

  • Getting started API keys, SDKs and your first session
  • Concepts Sessions, verbs, sovereignty and provenance
  • Guides Workflows, credentials, triggers, pools and live view
  • Reference Every endpoint, error and limit
  • Benchmark DACH-Web-Agent-Bench, open enough to rerun yourself