Every page you'll advertise, graded — yours and your rivals'.
Before a dollar of spend, GooseCabbage crawls your whole site — and every known competitor's — and grades each page on the job it's actually there to do: a homepage sells positioning, a pricing page makes cost legible, a sign-up page removes every reason not to convert.
Then it does something no other tool does: it scores how ready each page is for the AI agents that now browse and act on the web — structured data, semantic markup, and WebMCP tools included.
- JSON-LD structured data
- Semantic landmarks (4/4)
- Image alt-text · 61%
- No WebMCP tools registered
Your site and your competitors', on the same rubric.
Click through a real audit: per-page grades with the exact positives, negatives, and CTAs the agent surfaces, a head-to-head matrix against the field, and a live AI-readiness score you can watch climb.
- +Monthly / annual toggle
- +Money-back guarantee stated up front
- −Enterprise tier hides price behind "contact us"
- −No feature-by-plan comparison matrix
- −Two competing CTAs split visitor intent
Publish a feature-by-plan matrix and collapse to one primary CTA — Rival A already does both.
The best campaign in the world dies on a bad page.
Ad platforms grade your landing pages and quietly tax weak ones with higher costs. Meanwhile your competitors' pages are public — and a goldmine — but nobody has time to read them all. And a new audience has arrived: AI agents that browse, compare, and buy, and skip pages they can't parse.
You pay the landing-page tax.
Quality Score and ad rank fold in landing-page experience. A slow, vague, or untrusted page means you bid more for the same click — forever.
Competitor pages go unread.
Your rivals tell you their positioning, offers, and CTAs on every page. Reading all of them by hand is impossible, so you fly blind into the same auction.
Agents can't parse you.
AI shopping and research agents skip pages without structured data, semantic markup, or actionable tools. Invisible to agents means invisible to a fast-growing slice of demand.
Four lenses on every page.
Each page is graded only on the dimensions that matter for its type — then on AI-friendliness, every time.
Conversion fitness
Message clarity, CTA strength, form simplicity, distraction-free flows, cost legibility — scored against what the page type is for, 1–10, with concrete positives and negatives.
Trust & proof
Social proof, trust microcopy, guarantees, and risk flags that scare off high-intent buyers. The agent quotes the actual phrases on the page, not generic advice.
SEO & structured data
JSON-LD types, OpenGraph, meta descriptions, a sane heading outline, and semantic landmarks — the machine-readable layer that search engines and AI crawlers depend on.
AI-friendliness (WebMCP)
A deterministic 0–100 readiness score: WebMCP tool registrations, labelled actionable forms, alt-text coverage, and HTTPS — how easily an AI agent can understand and act on the page.
Crawl, grade, benchmark, route.
Crawl both sides
Grade by page type
Benchmark head-to-head
Route every segment
Built for the agentic web.
WebMCP (Web Model Context Protocol — Google/Microsoft, W3C draft) lets a page expose tools an AI agent can call. We detect both imperative registrations and declarative form annotations, then score readiness deterministically — no LLM guesswork, no tokens burned, accurate today and as adoption grows.
Imperative + declarative detection
We scan for navigator.modelContext.registerTool() and for <form toolname="…"> annotations, surfacing the actual tool names an agent could call on the page.
Deterministic, token-free score
Readiness is computed from extracted HTML signals, weighted toward WebMCP but crediting the agent-readiness fundamentals it's built on — so the badge is meaningful before WebMCP adoption is universal.
Foundations that compound
JSON-LD, semantic landmarks, a single clean H1, labelled forms, alt-text, HTTPS — the same signals lift you with search engines and with agents at once.
Levels you can act on
Every page lands at none, partial, or full readiness with the exact detected signals listed — so you know precisely what to ship next.