Vlon.ai

Turning a creative idea into production-ready textile artwork is slow, manual, and format-hostile — designers juggle PPT, PDF, and Keynote sources and hand-convert to SVG, PSD, and TIFF for screen, embroidery, and digital print. Each production method has different constraints, and a single edit means redoing the whole chain.

1. 01

   Magic Fill for colour and motif edits

   I built Magic Fill — non-destructive recolour and motif swaps rendered through a canvas + SVG (resvg) pipeline — so a single artwork adapts to a whole palette or collection without restarting from scratch.

2. 02

   AI generation & upscaling

   Generation and upscaling run through a ComfyUI pipeline plus fal.ai and Google Gemini models, tuned per target production method — screen, embroidery, or digital — and exported to real production assets (SVG/PSD via ag-psd, TIFF, packaged with archiver).

3. 03

   Git-style version control

   Designs get version history on a Drizzle-backed store, so teams branch, compare, and roll back artwork the way engineers do with code — critical when a print run depends on the exact right revision.

10×

Faster idea-to-print

3

Production methods supported

SVG/PSD/TIFF

Production export formats

Next.js · Express · Drizzle · ComfyUI · fal.ai · Gemini · Canvas · resvg