Should I vibe code
Create typography-heavy graphics, labels, posters, and merchandise-ready exports
The canvas is a weekend. The font licence you embedded in that merch export is a letter from someone's lawyer.
?
Their verdict, the Pro 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
You will get a canvas, draggable text, a few warps and an export in a weekend, and it will look like most of a design tool. It is not. Kittl's whole reason to exist is typography — text on a path with real kerning, OpenType alternates, layered vintage effects — and each of those is weeks, not evenings, before it stops looking almost right. The second half is not code at all. What the subscription actually covers is a licensed asset library and a commercial licence that says you may put those fonts and illustrations on things you sell. Your own build has no such licence, and the moment an export leaves the screen and becomes five hundred printed shirts, a missing font grant stops being a technicality.
What actually breaks
not "if". the specific failures.
- Text on a path, which is where every home-made design tool stops — kerning along a curve, baseline offsets and glyph rotation are the feature, not a nice-to-have
- Font licensing, which distinguishes viewing, embedding in an exported file, and putting the result on merchandise for sale, and grants those separately — a desktop licence covers none of the last two
- Export fidelity: a browser canvas is RGB at screen resolution, and a printer wants CMYK, 300 DPI, bleed and outlined text, so the file that looked perfect comes back muddy
- SVG filters and blend modes, which render one way in Chrome and another in Illustrator, so the shop's proof does not match your screen
- Undo across a grouped transform, which is the bug that makes people quit a tool they otherwise like
- The asset library — templates, illustrations, hundreds of licensed display faces — which is not code and cannot be generated
Is that you?
the verdict is a default, not a law
- Every font in it is one you own outright or that carries an OFL licence, and you have read what it says about embedding
- The output is screen-only: a thumbnail, a social image, something you post and forget
- You are automating one repetitive layout rather than building a general editor
- The design work is yours and nothing produced by it is sold
- Anything it exports gets printed on something you sell
- Clients use it to produce artwork whose licensing is then implicitly your problem
- You are scraping fonts, templates or illustrations from anywhere to fill a library
- You need CMYK, spot colours or a print shop's preflight to pass
If you build it anyway
the checklist, then the prompt that enforces it
- Keep a licence field on every font and asset, filled in before it can be used, and refuse the export if any layer references an asset with no recorded grant. That one constraint is most of the risk gone.
- Do not implement a vector engine. Build on an existing SVG or canvas library and let it own boolean ops, path maths and text layout.
- Convert text to outlines on export, and keep the editable version separately — it sidesteps a whole category of font-embedding argument and of renderer mismatch.
- Treat print export as a separate, tested path: CMYK, 300 DPI, bleed and a preflight check, verified against one actual proof before anyone orders a run.
- Use the browser's own undo model or a proper command stack from the start. Retrofitting undo onto a canvas app is a rewrite.
- Save documents as SVG or another format something else can open, not as a JSON blob only your renderer understands.
I am building a graphic design editor for typography-heavy artwork — posters, labels, merch. Two things matter more than features: I must not embed assets I have no licence for, and the file that goes to a printer must match what I see. Order the work that way and push back when I skip it. 1. Before any drawing code, build the asset registry. Every font and illustration carries its source, its licence, and explicit flags for embedding and for commercial or merchandise use. Nothing renders unless those fields are filled in. 2. Export refuses to run if any layer references an asset without an embedding grant, and names the offending layer. Do not add an override flag — if I ask for one, tell me why this check exists. 3. Never fetch fonts or artwork from a source I have not explicitly listed. Do not scrape a font site, and do not bundle anything from a template pack. 4. Build on an existing SVG or canvas library for path maths, boolean operations and text layout. Do not write a vector engine or a shaping engine. 5. Documents save as SVG or another open format with a small sidecar for editor state. No proprietary JSON blob that only this app can open. 6. Implement a command stack with undo and redo before the second tool. Retrofitting undo onto a canvas editor is a rewrite, especially across grouped transforms. 7. Print export is a separate, tested path: outline all text, CMYK, 300 DPI, bleed and margin, plus a preflight report. Tell me to order a single physical proof before any real run. 8. Screen export stays RGB and never claims to be print-ready. 9. Out of scope and say so rather than half-building it: text on a path, layered vintage effects, real-time collaboration and an asset library. Those are the product. If I want them, say that $15 a month buys the licences as well as the software.
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
The moment the output gets sold or printed. Around $15 a month is not really buying an editor — Penpot and Inkscape are free and better at the drawing — it is buying a licensed font and illustration library plus a commercial grant covering the things you make with it. That is the part with no open-source substitute and the part a court cares about.
$15/mo is cheaper than your weekend.
Save everything as SVG. If the documents are SVG plus a small sidecar for editor state, Inkscape, Illustrator, Figma and Penpot all open your work the day the project dies, and the only thing you lose is your own UI. Keep the asset licence registry as a file too — six months later, "which font is on that poster and am I allowed to sell it" is a question you will actually need answered.
Mature open-source collaborative design platform with SVG-native editing.
Decades-old open-source vector editor that already does text on a path and print-grade export.
Questions
Is the font licensing really a risk for a side project?
Only when money appears. Designing with a font you licensed for desktop use is fine; embedding it in a file you distribute is a separate grant in most licences, and putting it on goods you sell is a third. Foundries do check marketplaces, and the demand letter goes to whoever sold the shirt. Keeping a licence field on every asset costs an hour and removes the whole category.
Why so pessimistic when a canvas editor is a well-trodden weekend?
Because the weekend gets you rectangles and drag-to-move, and Kittl's users came for text on an arc with real kerning, alternate glyphs and a layered print effect. The distance between those two is where every hobby design tool stalls — not because it is impossible, but because it is months of typography work that never demos better than the thing you built on Saturday.
What is the one useful thing to build here?
The repetitive layout you personally do over and over — a label generator, a batch of variants from a spreadsheet, a template that renders forty product images. Automating one specific job with SVG and a script is genuinely better than a general editor, and it never runs into the parts of this page that hurt.
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.
The editor is hard, the template library is harder, and the font licensing is a legal department.
You can build the canvas in a weekend. You cannot build a licence to 170 million stock assets for ten dollars.
You can rebuild the editor in a weekend. You cannot rebuild the licence on five million stock photos.
last reviewed 2026-08-04 · verdict is editorial and unsponsored · shared entry data from canivibecodeit under MIT · not legal advice