security / browser scraping

Browser And Scraping Safety

Browser and scraping docs explain how Ambient exposes powerful web capabilities without treating arbitrary pages as trusted code.

Developer Previewllms.txtSource evidence
Capability Sandboxing Model Ambient prefers sandboxing risky capabilities so the agent can operate on a regular desktop with bounded authority. Capability Sandboxing Model Ambient prefers sandboxing risky capabilities so the agent can operate on a regular desktop with bounded authority. classifies riskapplies policycontains executionproves outcome Sandbox boundary policyisolationaudit 1 Untrusted surface MCP serverPi packageBrowser pageScraper 2 Ambient mediation Descriptor reviewPermission policySecret redactionToolHive 3 Execution boundary Container workloadSandboxed Pi toolURL policyAudit log 4 User proof Prompted approvalVisible outputDiagnosticsRollback
Ambient prefers sandboxing risky capabilities so the agent can operate on a regular desktop with bounded authority.
Product screenshots

Main shell

The main Ambient Desktop shell combines chat, workspace context, navigation, and evidence panels in one local developer workstation.

Ambient Desktop main shell with chat, sidebar, and workspace controls.

Threat Model

Autonomous agents can load pages a user would never inspect manually. Browser and scraper capabilities use URL policy, private-address protection, isolation, and user approval for sensitive actions.

Safe Defaults

  • Prefer vetted browser/scraper providers for core documentation workflows.
  • Treat local/private network access as a separate approval boundary.
  • Keep full large outputs as artifacts with bounded previews.
  • Show screenshots and logs as evidence without leaking secrets.