Renda Converts Claude Design HTML Zips into Instagram/Twitter PNGs
HTML-to-PNG for social posts when Canva and Figma already export directly.

Image-to-HTML using vision models first, but v0.dev already dominates this space.
Frontend developers, designers
v0.dev · Screenshot-to-Code
I have been trying to get something like this working in one form or another for close to a year. The latest image models finally make it possible.
It's still early, but I'm really interested in this approach to AI UI - using image models before coding models for front end development. The challenges are; 1 - How do we get image models to generate decent UI 2 - How do we convert those images into code But I think I've got something which tackles this well enough to be useful, at least in my workflows.
Would love feedback and if this is worth continuing to explore!
HTML-to-PNG for social posts when Canva and Figma already export directly.
CPU-only VLM OCR beats Tesseract on layout without needing CUDA or cloud APIs.
CPU-only VLM OCR beats Tesseract accuracy without sending data to the cloud.
Parallel model generation is nice, but Stripo and Beefree already solve this.
No-signup image-to-3D generator that exports OBJ and GLB instantly.
Tool-calling on 3B parameter models offline bridges local toys and actually useful assistants.