Public pages only
The scan follows same-origin public links and reads content available without an account.
Product
Preview an AI-agent website audit.
See the evidence, task outcome and performance signals planned for the public domain scan before it opens.
Useful evidence before registration.
The public demo is designed to return meaningful findings without requesting credentials or changing the target website.
One safe agent task
A browser executor identifies the offering and reaches the primary next step without submitting a form.
Three verified insights
Each result includes an observation, evidence reference, impact and a concrete remediation.
What the free result will explain.
The report connects a business goal to observable steps and a verified terminal state. It does not reduce the site to one unexplained score.
Agent task
Illustrative sample data
The rendered journey
The exact viewport, target and final state available to the browser agent.
The semantic interface
Names, roles and relationships used to interpret the same page.
Dependencies behind the result
Requests, waits and runtime signals connected to the observed task state.
- Goal
- Understand the product and find the primary next step
- Result
- Success with friction
- Verified outcome
- The offering and waitlist were found. One interactive target moved during the run.
- Evidence
- Navigation trace, before-and-after screenshots and an accessibility tree excerpt.
Works well
Primary action is machine-readable
The main CTA is a semantic link with an accessible name and a stable destination.
Keep its purpose and destination consistent across locales.
SEM-004Needs attention
A late banner shifts the target
The CTA moves after the agent has identified it, increasing the chance of a wrong action.
Reserve banner space before the page becomes interactive.
PRF-011Works well
Structured facts match visible copy
The page purpose and organization name agree across visible content and JSON-LD.
Re-test parity whenever key marketing claims change.
DAT-003Speed and stability stay attached to the task.
A fast page can still be difficult for an agent. The preview reports completion time, observable steps, retries and layout movement together.
- Task elapsed
- 18.4 s
- From goal delivery to the independently verified outcome.
- Observable steps
- 7
- Browser actions visible in the session trace, excluding hidden reasoning.
- Retries
- 1
- Repeated actions are separated from the successful path.
- CLS
- 0.18
- Layout movement measured during the same scenario run.
The preview shows its limits.
Free results separate what was observed from what still requires ownership verification, a test account or a sandbox.
Visible without signup
- Public crawl and rendered-page coverage
- One vendor-neutral browser task
- Three evidence-backed insights
- Task timing, steps, retries and CLS
- A concise technology snapshot
Not inferred from a public scan
- Authenticated account journeys
- Checkout, booking or payment completion
- Cross-agent repeatability claims
- Private API, OpenAPI or MCP behavior
- Security posture beyond observed public signals
What changes when the scanner opens.
This page becomes an interactive domain audit only after the runner can return real events, preserve scan state and enforce public-safety boundaries.
- A domain input with SSRF-safe target validation
- Real stage-by-stage progress and cancellation
- A meaningful result before any registration gate
- A signed claim flow that keeps the completed scan
Get access when the domain audit opens.
Join the waitlist for the public preview, or apply as a launch partner to test a real workflow in a controlled environment.