# Drawlab

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Baptiste Borel](https://cerebralvalley.ai/u/bborel-dev)
- **GitHub:** https://github.com/bborel-dev/drawlab
- **Demo video:** https://www.youtube.com/watch?v=7eI62Y7cO3s
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2916

Problem: AI image tools give you flat pixels, and classic vectorizers give you hundreds of anonymous shapes. Designers can't edit either one.

Solution: Drawlab doesn't generate, it draws. Drop an image and AI agents redraw it live, stroke by stroke, into a clean SVG with one named, editable layer per element.

How: Muse Spark plans every element, SAM 3.1 cuts each one out, Muse Image rebuilds hidden parts, and a vector engine draws the layers. Then you edit by hand or just ask for a change in plain words.

Open source, built solo during the hackathon.

## More from Meta Global AI Developer Hackathon

- [CLIQ](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2913)
- [IsOneGuyConsideredAteam](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2914)
- [sksaddam](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2915)
- [BELIEVE](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2917)
- [prakhar](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2918)
- [Croncrete](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2919)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/2916. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt