KAGEWebsite Intelligence
Daily quota spent

KAGE has judged enough for today

Today's budget for new analyses is gone. Your site is spared, for now.

A new round starts tomorrow. Nothing was charged to you.

  • Read the analyses that are already finished.
  • See how KAGE hands down a verdict.
  • Come back tomorrow and put your own site under the lamp.
Read the verdicts ?Every analysis runs a real browser, a Lighthouse pass and a model that reads your page. That costs real money each time, so a daily ceiling is what keeps the tool free.
KAGEWebsite Intelligence
In the works

The full teardown

The free analysis is already the real thing, run on your page. The teardown goes further.

  • Fourteen sections, each scored and argued from your own captures.
  • Ten priorities in order, with the effort and the effect for each.
  • Your hero rewritten, plus a recommended page structure.
  • Three moves for the business, not only for the page.
Run the free analysis
Two free analyses. No account.

People judge your website in seconds.
KAGE shows you why.

Paste a URL. You get the price a first-time visitor puts on your page, what is stopping it from winning work, and where it really stands on speed, visibility and accessibility on a phone.

Perceived valueSpeedSearch visibilityAccessibilityTrust

One for your site. Use the second on a site your buyers compare you with.

What you get

Real websites, from the cheapest looking to the most expensive.

Not mockups. Sites you can open yourself, run through the same analysis you get. Each one carries the date it was captured.

desktop capture
    53%

    of mobile visitors leave a page that takes longer than three seconds to load. Slow, outdated and invisible each carry a price, and nobody tells you they left.

    Google, mobile speed benchmarks, 2017
    +8.4%

    more retail conversions from a site that is one tenth of a second faster on mobile. Travel: +10.1%.

    Google and Deloitte, Milliseconds Make Millions, 2020
    46%

    of what people said about a website's credibility was about how it looked, before anything it said.

    Stanford Persuasive Technology Lab, 2,684 participants, 2003
    34.5%

    fewer clicks on the first search result when Google shows an AI Overview above it.

    Ahrefs, 300,000 keywords, 2025
    What is measured

    Technical checks and reasoned recommendations, with the source named on each one.

    Every finding says what it rests on: a measurement, a documented practice, or a designer reading this page. A number that carries a limit says so next to the number.

    • 01Nielsen Norman Group

      Ten usability heuristics, reading patterns, form and navigation research.

    • 02Google Core Web Vitals

      Performance diagnostics from Lighthouse, plus Core Web Vitals from real visits where Google has them.

    • 03WCAG 2.2

      Contrast, tap targets, alt text, labels and structure at level AA.

    • 04Google Search Essentials

      Titles, descriptions, indexability, structured data, mobile-first indexing.

    • 05Baymard Institute

      Two decades of conversion research on what makes people trust a page.

    • 06Stanford Web Credibility

      How visitors decide in seconds whether a site is credible, and what design has to do with it.

    What we judge, and what we leave alone

    Not everything deserves a score.

    Your audience and your goal decide most of it. One layer is not up for debate, and that is the layer we measure.

    MeasuredNot an opinionPublished standards, with numbers. Either the page passes or it does not.
    • WCAG 2.2 AA: 4.5:1 contrast on body text, visible focus, 24x24 px tap targets, labelled fields.
    • Core Web Vitals: LCP under 2.5 s and CLS under 0.1 from a Lighthouse run, INP under 200 ms from real visits where Google has them.
    • Google Search Essentials: indexable page, title and description, usable on a phone.
    • Mechanics: nothing overflows at 320 px, links go where they say, images carry alt text.
    ConventionEvidence, not lawYou may break these. It should be a decision, not an accident.
    • One clear next step per screen.
    • Line length around 50 to 75 characters, body text 16 px or more on a phone.
    • Headings in order, H1 then H2, not picked by font size.
    • Nielsen Norman heuristics: visible system status, consistency, recognition over recall.
    TasteWe do not score thisYour niche, your audience and your goal decide it, not a checklist.
    • Palette, typeface, photography style.
    • How loud or how quiet, how much copy belongs on the page.
    • Whether the price is shown, and how.
    • A law firm and a tattoo studio cannot look the same, and neither one is wrong.

    So every finding is labelled. Measured: a published threshold, met or not. Convention: a researched practice you may break on purpose. Judgement: the model reading your page in context, and you are free to reject it. Taste gets no label, because it gets no score.

    The competitor

    You·
    vs
    Them·

    Same market. Same judge. Very different first impression.

    Your customer is comparing you anyway. This puts the comparison on screen, and both pages sell to the same buyer.

      Compare two sites
      Pricing

      The verdict is free. The plan to fix it is paid.

      One-time prices, no subscription.

      The analysis
      €0Free
      Two analyses
      Analyze a site
      • Price, verdict and KAGE score
      • Flags on your own screenshot
      • Speed, search and accessibility readings
      • The priorities this reading actually found
      • A card you can post
      Full teardown Most useful
      €49EUR
      One-time
      • Every issue, explained
      • Fourteen sections, desktop and phone
      • Copy, buttons, type, color, images
      • Search and AI visibility
      • What the business should lead with
      • Ten priorities, in order
      • It comes off the price if we build together
      • Free for clients of Susakovich, by access code
      Prefer a personDesigner teardown, €149. Redesign with Susakovich from €1,900.

      Danijel reviews your page on video, on your actual site, with the fixes in order. The studio builds the new one.

      Talk to the studio
      Questions

      Before you paste.

      01Is the price real?

      No. It is what the page looks like it cost to a first-time visitor, a perceived-value heuristic. It matters because that is the judgment your clients make, in seconds, before reading.

      02Why not just ask ChatGPT?

      For a first opinion, do. The difference is what the answer rests on. A chatbot reads what you paste into it. KAGE opens your page in a real browser at desktop and phone size, takes both screenshots, reads contrast ratios and tap-target sizes out of the live page, runs a Lighthouse pass, and takes Core Web Vitals from Google’s real-visitor data where your page has enough of it. Those numbers are measured, not remembered. And every run uses the same brief and the same fourteen categories, so your page and your competitor’s are comparable, and so is the same page a month later. Ask a chatbot twice and you get two answers, both confident.

      03Do you store my site?

      The reading is cached for 24 hours, a full teardown for 7 days, and a log line for each analysis is kept with a one way hash instead of your IP. Nothing is published, there is no account and no cookie, and you are only emailed if you ask Danijel to write back.

      What KAGE keeps, line by line

      04Why only two?

      Each analysis runs a real browser, a Lighthouse pass and a real model, and that costs real money on every run. Two is the useful pair: your site, and one your buyers compare you with. After that the teardown is the paid product, and looking again at a site already analyzed is always free.

      05Can it be wrong?

      Yes. Measured values come from the page itself: what is read out of the code, like contrast or tap-target size, repeats exactly, while a Lighthouse performance score moves a little between runs. The verdict is a briefed model’s opinion. Treat it like a sharp first review, then get a person for the decisions that cost money.

      06Who is behind this?

      Susakovich, a web, brand and automation studio in Munich. KAGE is the diagnosis. Susakovich builds.

      07Why KAGE?

      KAGE (影) means shadow in Japanese. The idea is that the tool surfaces what usually goes unnoticed on a site. There is no grand meaning behind the Japanese: the name sounded right and the concept happened to match.

      Unlock the full teardown, €49