Available

Know what an agent can do with your site.

Scanner reads a website the way an AI agent reads it, and scores what it finds out of 100.

Available by request, not self serve. Send a domain and we run the scan and send the report back. Fifteen sample reports are already published.

Scan your site See a sample report

01Why it matters

An agent that cannot book you books someone else.

The report names the thing in the way on each page: the CAPTCHA, the unlabelled field, the missing catalogue.

A scan, and what comes back

npm run scan -- https://yourbusiness.com.au --json report.json

{
  "total": 48,
  "grade": "partially ready",
  "categories": {
    "crawlability": 12,
    "structured_data": 0,
    "action_affordances": 14,
    "form_quality": 16,
    "agent_interface": 6
  },
  "agentCallableActions": 0
}

Trimmed from a real report in the published sample. No business is named.

02How it works

01
Send a domain
No account, no snippet, nothing to install.
02
We score five things
Crawlability and discovery, structured data, action affordances, form quality, agent interface. Twenty points each.
03
You get a shareable report
The score, the blockers named one by one, and the fix for each.

03In detail

What the five categories reward

The five scoring categories, what each rewards, and what each is worth.
Category	What it rewards	Worth
Crawlability and discovery	robots.txt posture toward agents, sitemap, llms.txt, an ARD catalogue, canonical	20
Structured data	JSON-LD, business entity type, opening hours, phone and address, services	20
Action affordances	Booking, quote and contact forms, known booking widgets, real buttons rather than clickable divs	20
Form quality	Labels, name attributes, input types, a real submit control, and a CAPTCHA penalty	20
Agent interface	WebMCP signals, a persistent MCP endpoint, semantic landmarks, alt text	20

Grades: 70 or more is agent-ready, 40 to 69 partially ready, under 40 invisible to agents.

04Evidence and limits

45.7

Mean score out of 100, across fifteen Sydney service sites

1 of 15

Sites with an action an agent could actually complete

0 of 15

Sites reaching the agent-ready grade of 70. The best scored 69.

Source: packages/scanner/benchmark/reports/summary.json · 25 Aug 2026

Send us a domain.

We will tell you what an agent can and cannot do with it, free.

Scan your site See a sample report

Know what an agent can do with your site.

Send a domain

We score five things

You get a shareable report