Should I vibe code
Design-focused website publishing with flexible layouts and type tools
Cargo's $19 is a foundry licence with a website attached. Copying the woff2 into /public is not the same deal.
?
Their verdict, the Cargo (single plan) price and the build-time estimate come from their entry, MIT-licensed. Checked 2026-08-04.
?
Our verdict, the regret score and everything below it. Editorial and unsponsored — nobody can pay to be moved.
The honest answer
why the verdict is what it is
Read Cargo's font page and the entry writes itself. The $19 is not really buying hosting — hosting is a solved commodity and your agent will hand you a better-performing site than Cargo's in an afternoon. It is buying a webfont licence to the entire Dinamo library, cleared for every site on the account, which is why designers pick it and which is the one component of this product you cannot reproduce by writing code. Everything else you can: the grid, the type controls, the publish button. What you would be signing up for instead is the other half of Cargo, which is a real-time collaborative design canvas with pixel-accurate output across browsers — the single hardest UI genre there is, and the reason every builder in this category has a decade of engineering in it. Build the site. Do not build the builder, and do not copy the woff2.
What actually breaks
not "if". the specific failures.
- Fonts, legally. A desktop licence lets you set type in Illustrator; it does not let you serve the file from your own domain. Self-hosting a commercial typeface without a webfont licence is a licensing breach, foundries do audit, and the invoice arrives as a retroactive licence at their price rather than a takedown
- Fonts, technically. Variable axes, OpenType features, real small caps, hinting at 11px on Windows, and the FOUT that makes a typographic site look broken for 400ms on every cold load
- The canvas. Free-position layout that survives a 320-pixel phone is the problem every builder in this category has spent years on, and an agent will give you absolute positioning and a very confident demo
- Cross-browser text rendering, which is where design-led sites actually die: the same stack sits differently in Safari and Chrome, and your client uses the one you did not test
- Real-time multi-user editing, if you promised it. Conflict resolution on a design document is CRDT territory and it is not a weekend
- Commerce, the moment you add the shop. Now there is checkout, tax, fulfilment status and refunds, and this entry's risk numbers stop applying — read the payments entries instead
- Custom domains at scale: DNS instructions for people who have never seen a CNAME, certificate issuance and renewal, and the apex-domain problem that has no clean answer
- Your own uptime expectations. Artists send this link to galleries and clients; a site that is down on the day of a submission deadline is a specific, dateable loss
Is that you?
the verdict is a default, not a law
- It is your own site, using typefaces you have a webfont licence for or open ones from a foundry that grants it — and you have read the licence rather than assumed
- The output is static HTML and CSS you could hand to any host, with the content in files you can read
- There is one editor of the site and that editor is you, so the collaborative canvas never has to exist
- No shop. The instant money changes hands this is a different entry with different numbers
- You are self-hosting a commercial typeface because it was already in your Fonts folder — that is the failure mode this entry exists to name
- Other people are publishing through it, because now you own their downtime and their DNS support requests
- You have promised anyone a visual editor with free positioning that also looks right on a phone
- The commerce add-on is what you actually wanted, in which case the honest comparison is $9 a month against building a checkout
If you build it anyway
the checklist, then the prompt that enforces it
- Settle the font licences before you write a line of CSS. Check whether each typeface has a webfont licence, what page-view band it covers, and whether self-hosting is permitted at all. Foundries publish this and it takes ten minutes.
- Prefer open licences — SIL OFL faces are genuinely excellent now — or a hosted foundry service where the licence is the subscription and the fonts are not your problem.
- Serve fonts with font-display: swap, preload the one face above the fold, and subset aggressively. A typographic site with a 900ms flash of fallback text has failed at its only job.
- Design in a constrained system rather than free positioning: a grid with named breakpoints beats absolute coordinates that need re-solving at every width.
- Keep the output static and the content in flat files, so the site outlives whatever framework you used to generate it.
- Automate certificates and set an expiry alarm you actually receive. The dominant failure in this whole category is a lapsed certificate on a site nobody visits daily.
- If commerce ever appears, use a hosted checkout and never let card details touch your origin. That single decision keeps this entry's risk numbers roughly true.
I am building a design-led portfolio and publishing tool: strong typography,
flexible layout, custom domain. The risk that is not obvious is licensing —
the fonts that make this look good are usually not mine to serve. Work in this
order and stop me if I get impatient.
1. Before any styling, list every typeface I want and, for each, tell me
whether I have a webfont licence, what it covers, and whether self-hosting
is allowed. If I cannot show one, propose an SIL OFL alternative instead.
2. Never copy a font file out of my system Fonts folder into the project. If I
ask you to, refuse and explain what a desktop licence does and does not
permit.
3. Build the type system next: modular scale, line lengths, real small caps and
OpenType features where the face supports them. Subset the fonts to the
glyphs actually used and preload only the above-the-fold face.
4. Use font-display: swap and show me the cold-load behaviour on a throttled
connection before we talk about layout.
5. Layout is a named grid with explicit breakpoints. No absolute positioning
for content. Every page must be checked at 320px before it is called done.
6. Output is static HTML, CSS and images from flat source files. No database,
no server rendering, no lock-in to the tool that built it.
7. Custom domain: automated certificate issuance with renewal, an alarm that
reaches me on failure, and written DNS instructions I could follow at 60.
8. Test the same page in Safari and Chrome side by side and show me the
differences in text rendering. Do not tell me they are identical.
9. Do not build a visual editor, a canvas, or real-time collaboration. If I ask
for one, tell me that is the multi-year part of Cargo and I am paying $19 a
month to skip it.
10. Commerce is out of scope. If I want a shop, the answer is a hosted checkout
that never touches my origin, and a different conversation about risk.That one keeps you out of trouble. For the prompt that actually builds it, canivibecodeit.com has one.
their build prompt ↗Or don’t build it
the boring option, and the way back out
If typography is the point. Nineteen dollars a month bundles a foundry library you would otherwise licence separately, plus hosting, domains and certificates, which makes it one of the better-value products on this site for exactly one audience — designers who want to set type properly and never think about a build pipeline. If you have your own licences and want a static site, build it and enjoy it. If you were about to self-host a typeface you have not licensed, the subscription is cheaper than the letter.
$19/mo is cheaper than your weekend.
Clean if you stayed static: flat source files in, HTML out, any host will take it, and the domain was always yours. Two things do not travel. The fonts are licensed to a context — moving hosts is fine, changing traffic band or handing the files to a client may not be, so keep the licence PDFs beside the repo. And anything that lived only in a visual editor's saved state is a proprietary format of one, so if you ever do build a canvas, make its serialisation something you would be willing to read by hand.
Open-source visual website builder — the closest honest comparison to the half of Cargo that is a canvas.
Embeddable page-builder framework, worth reading before writing your own editor from scratch.
Questions
Is font licensing genuinely a real risk, or is this scaremongering?
It is real and it is routine. Webfont licences are sold separately from desktop licences, usually banded by page views, and foundries do run automated checks against public sites for their own files. The outcome is not normally a lawsuit — it is a bill for the licence you should have bought, backdated, at list price. Cargo's own font page is the tell: they licensed the Dinamo library so their subscribers do not have to, and that arrangement is a large part of what the subscription is.
Carbonmade is in the same category. Which entry should I read?
They fail differently. Carbonmade's problem is the image pipeline — four hundred large photographs, colour management, bandwidth. Cargo's is the editor and the type: a free-positioning canvas that has to survive a phone, and a font library that is licensed rather than written. If you are a photographer, read that one; if you are a designer who cares which grotesk it is, this one.
What if I only want the layout tools for my own site?
Then this is a good weekend. A constrained grid, a proper type scale and static output will get you most of the visual result with none of the risk, and the whole thing stays a folder of files. The verdict is DEMO ONLY because the product is the editor, not the site — and the editor is the part you will abandon in week three.
Every week, someone ships something they shouldn’t have.
New verdicts, the worst thing that landed in the trap, and the occasional incident report. No other email, ever.
Nobody notices your layout. A photographer notices the second your resizer throws away the colour profile.
Nobody wants a website builder. They want to change the opening hours on a Sunday without texting you.
A text-first site builder is a static site generator with a nicer front door. You already own three of those.
last reviewed 2026-08-04 · verdict is editorial and unsponsored · shared entry data from canivibecodeit under MIT · not legal advice