◆ facts, not opinions

X-ray any funnel
into facts.

Paste a competitor's URL. We walk their funnel to checkout and extract the exact design tokens, proof language, and conversion tactics they use — every claim backed by on-page evidence. No AI opinions.

Live scanning is in private beta — submit to browse 32 real teardowns.

◆ 32 funnels x-rayed · 164 tactics detected · 266 evidence strings · 100% deterministic

What 32 x-rays revealed

Everyone knows the playbook.
Almost nobody runs it.

We walked 32 funnels to checkout — AI tools, dev platforms, DTC — and logged every proof element, guarantee, and tactic against on-page evidence. The gaps below are measured, not surveyed.

91%let proof go quiet before checkout. Doubt peaks at the credit-card field — and their last testimonial is pages behind. Proof isn't a section. It's an escort.

56%make you scroll to find a reason to believe. Above the fold: headline, button, hope. The first proof point is below it — the back button isn't.

91%offer no money-back guarantee. The buyer carries 100% of the risk. A guarantee is one sentence — the cheapest tactic almost nobody deploys.

81%never run the us-vs-them comparison. If you don't stage it on your page, your buyer runs it in another tab — and your competitor writes the script.

Every number links to the exact string it was measured on. The only unmeasured funnel in this dataset is yours.

Already on the table

32 funnels, taken apart

Click any — its tokens, its tactics, and the exact strings that convicted it. Names you know included.

How it works

Paste a URL. Get the receipts.

Point it at a funnel

Any URL — we auto-map the funnel, or you define the exact flow step by step.

We walk it, step by step

Lander → product → cart → checkout. Each step screenshotted and measured on desktop and mobile.

Get the teardown

Design tokens, proof language, detected tactics — each with evidence — plus an agent-ready build brief.

What you get

Measured, not summarized

✓

Exact design tokensColors, fonts, type scale, radius — as W3C tokens + Tailwind vars.

✓

Proof-language templatesThe reusable social-proof formulas they deploy, with live examples.

✓

Detected tactics + evidencePricing, guarantees, urgency, logo walls — each tied to the string it fired on.

✓

Desktop ↔ mobile diffWhat the funnel hides or shows by device — measured.

✓

Agent-ready build briefThe whole teardown as one grounded spec — paste into Claude or Cursor.

{
  "$description": "Design tokens extracted from ryzesuperfoods.com funnel (2026-07-07). Roles curated; palette from computed CSS.",
  "color": {
    "primary": {
      "$value": "#ca5839",
      "$type": "color"
    },
    "accent": {
      "$value": "#f3d7c9",
      "$type": "color"
    },
    "bg": {
      "$value": "#fcfbf3",
      "$type": "color"

Why measured beats summarized

Facts you can verify and export

	UXXRAY	Manual audit	AI-summary tools
Exact design tokens (hex, font, radius)	✓	~6 hrs	✗
Proof-language templates	✓	✗	✗
Evidence string per claim	✓	partial	opinions
0 hallucinations	✓	✓	✗
Agent-ready build brief	✓	✗	partial

The 0-hallucination guarantee

Every claim carries its evidence string. If a fact isn't provably on the page, it doesn't ship. No scores, no opinions — just what's measurably there.

Questions

The honest answers

Is this legal?

We show extracted facts — colors, fonts, tactics, proof strings — which aren't copyrightable (the BuiltWith model). We don't re-host a competitor's content; screenshots are thumbnailed evidence.

Why no AI score?

LLM scores drift and hallucinate. We measure deterministically; every claim is backed by the exact element or string it came from.

What funnels work?

Standard e-commerce (Shopify/Woo) auto-map to checkout; affiliate/SaaS landers map to their key pages; or define the exact flow yourself, URL by URL.

What do I actually get?

Screenshots per step, a brand token kit, proof-language templates, detected tactics with evidence, desktop↔mobile diffs, and a one-click agent build brief.

X-ray any funnelinto facts.

Everyone knows the playbook.Almost nobody runs it.