Worked examples figma.com
Is figma.com vibe coded?
Design tool, designed. We fetched the homepage on 29 August 2026 and ran all 86 detectors over the raw HTML — the same scan you get from the box on the front page, frozen here so the numbers can be argued with.
www.figma.com
Hand-written
Somebody sat down and made decisions. Several of them were opinions.
- Vibe tells
- 6
- Craft tells
- 9
- Confidence
- high
Plenty of signal to work with.
Where the score came from
- Platform fingerprintsnone
Artefacts of the tool, not of taste
- Colour, type and surfacenone
The visual defaults, applied unedited
- Layout and architecture1 hit
The shape a page takes when nobody asks what it needs to say
- Copy tells2 hits
How the words were written
- Leftoversnone
Scaffolding never cleaned up
- Source and build1 hit
What is visible in the shipped code
- How it behaves2 hits
Weight, mobile, accessibility, SEO — where the money leaks
- Human craft9 hits
Evidence of deliberate decisions
Evidence
15 findings- +10
Heavy for what it renders
27 script files and 1630 KB of HTML to render 598 words. A full framework, an icon library and an animation library for a page a hundred kilobytes could serve.
- +9
Copy that never says who it is for
598 words, and still no price, no contact route and no named audience. The page describes a category, not a business.
- −9
Deliberate social preview
Open Graph title and image were both authored.
- −9
Skip-to-content link
A keyboard skip link exists. Generators do not do this.
- +8
Dependency sprawl
1 third-party asset hosts and 27 script files. Each sub-problem got whatever library was most popular in the training data, and nothing was ever consolidated.
- cdn.sanity.io
- +8
Fixed pixel widths that will overflow a phone
Hard pixel widths wider than a phone. Layouts break at sizes nobody prompted for, because desktop was the only viewport that got looked at.
- width:1440px
- min-width: 768px
- min-width: 960px
- min-width: 1920px
What we read
- title
- Figma: The collaborative canvas for design, code, and AI
- generator
- —
- server
- unknown
- status
- 200
- html
- 1.59 MB
- words
- 598
- fetched in
- 1919 ms
Reading this score
figma.com landed on 6/100, which puts it in the hand-written band. That number is not a percentage and not a probability — it is strength of evidence: 35 points of tells, damped by craft, less 62 points of deliberate decisions, mapped through one logistic curve.
What pushed it up
- Heavy for what it renders +10 — 27 script files and 1630 KB of HTML to render 598 words. A full framework, an icon library and an animation library for a page a hundred kilobytes could serve.
- Copy that never says who it is for +9 — 598 words, and still no price, no contact route and no named audience. The page describes a category, not a business.
- Dependency sprawl +8 — 1 third-party asset hosts and 27 script files. Each sub-problem got whatever library was most popular in the training data, and nothing was ever consolidated.
- Fixed pixel widths that will overflow a phone +8 — Hard pixel widths wider than a phone. Layouts break at sizes nobody prompted for, because desktop was the only viewport that got looked at.
- Everything centred +7 — 8 centred blocks. Long centred text is measurably harder to read, and with no left-aligned anchor the hierarchy flattens.
- The rule-of-three cadence +6 — 7 triplet lists and 0 “not just X — it's Y” constructions. Rhythm borrowed from marketing prose, with none of the content.
What pulled it down
- Deliberate social preview −9 — Open Graph title and image were both authored.
- Skip-to-content link −9 — A keyboard skip link exists. Generators do not do this.
- prefers-reduced-motion honoured −8 — Motion preferences are respected in inline styles.
- Hand-written ARIA −7 — 63 explicit accessibility attributes.
- No utility-class soup −7 — Not one Tailwind utility class in the markup.
- Structured data −6 — JSON-LD is present — somebody thought about search.
- Semantic landmarks −6 — 6 landmark elements instead of a pile of divs.
- CSS custom property system −6 — 186 CSS custom properties — a real token layer.
- Canonical URL −4 — Canonical link is set.
Craft we looked for and did not find
These are the actionable half of the report. None of them are hard, and every one of them would have moved the number.
What this scan cannot see
One request, no JavaScript executed. So everything below is invisible to us, and a page can be excellent or awful at all of it without moving the score:
- Domain registration age — needs WHOIS.
- Whether the nav's /about and /pricing actually resolve — needs a crawl, not one fetch.
- package.json dependency count, orphaned code, clashing abstractions, test coverage — needs the repo.
- Backend security shortcuts: string-built SQL, verify=False, missing auth on endpoints — needs the source.
- Whether row-level security or Firebase rules are actually switched on — needs the project.
- Console errors, hydration warnings and runtime 404s — needs a browser executing the page.
- Real Lighthouse numbers and horizontal scroll at 375px — needs a rendered viewport.
- Whether a JS-handled form posts anywhere — needs a submitted request.