I was working on a logo for another project and needed a usable SVG version.
Asking ChatGPT to recreate it as SVG produced a mess, and several online converters only revealed the account requirement or paid plan when I tried to export the result.
I eventually stopped looking and built the workflow I wanted: upload a PNG, JPEG or WebP, compare the raster and vector versions, adjust the number of colors and smoothing, and download the SVG.
It’s designed specifically for logos and flat-color graphics, so it won’t magically turn photographs or very complex artwork into perfect vectors. I’m still testing where that boundary is.
The project is here: https://vectorize.stemonte.io
I’d appreciate honest feedback on the actual output: are the generated paths clean enough to edit and use, or do they still need too much manual cleanup?