Paste your URL, pick a benchmark or name the sites you admire, and see exactly where your page's structure, copy and layout differ from theirs.
Free, no login, no email. Nothing stored beyond a 24 hour cache. Markdown copy
Give us your page, and a benchmark
Pick a curated pack — modern developer tools, or SaaS marketing pages — or name up to three sites you want to be measured against.
We measure both sides
Your page gets the full treatment: text, markup, a screenshot, and the type sizes and colours the browser actually resolved. The reference sites are measured for structure and copy.
Get the differences, with numbers
Not a score. A list of concrete differences — your hero is 61 words, the benchmark runs 21 to 38 — with what to change for each.
Everyone has a site they admire, and almost nobody can say what they admire about it. The instinct is to name the visible things — the colours, the typeface, the gradient — which are the parts that transfer worst. Copy those and you get a page that looks derivative and still does not work, because what made the original work was never its palette.
What actually transfers is structural. How many things the page claims to be. How long the hero runs before it stops. Whether one button is obviously the one to press. How many sections there are before the page ends. Those are countable, which means the difference between your page and theirs can be stated as a number rather than a feeling.
This tool does not grade you. A score would invite you to optimise it, and some of your differences are deliberate. What you get is the list with both numbers on it — your hero is sixty-one words, the benchmark runs twenty-one to thirty-eight — so you can decide which gaps you meant and which ones happened to you.
Every benchmark separates the patterns that transfer from the identity that does not, and the review is instructed never to recommend adopting a reference site's palette, typefaces, illustration style, or copy voice. If the honest read is that your own visual identity is stronger than the benchmark's, it will say so. A tool that turns every page into the same page is worse than no tool.
When you name a site, we read its markup rather than rendering it. That is enough to count its sections, its navigation, its hero length and its buttons, and it is not enough to see its colours or its type scale. Those fields come back marked not measured, and the review is forbidden from comparing against them. Curated packs were measured with a browser in advance, so they carry more. Either way, the side being reviewed — yours — is always measured properly.
No, and it is built specifically to avoid that. Every benchmark separates the patterns that transfer — how a page is organised, how much it says, how it asks for the click — from the identity that does not, meaning palette, typefaces, illustration style and copy voice. The tool is instructed never to recommend adopting those, and if your own visual identity is stronger than the benchmark's it will say so.
No login, no email, no ads. Same as this one.
Superflow agents review every page you ship against the standards your team holds, and leave what they find as comments on the page itself.