Scout in development · launch 1 Dec 2026. Try it live · docs

Use Case

AI Agents + Scout

Instead of giving every agent raw browser access, the agent calls Scout as a reliable check layer and gets structured results back.

Why this helps

  • One response shape instead of a tool mix per agent
  • Clear split: agent decisions vs browser evidence
  • One key and credit model across many agent workflows

Example pattern

typescript
// agent asks for factual page check
const result = await scout.assert({
  url: "https://example.com/pricing",
  assert: [{ selector: "#buy", visible: true }]
})

if (!result.passed) {
  // escalate to human or retry flow
}