# vme

- **Event:** [Hackathon: Build Your AI Teammate with CodeRabbit & Cline](https://cerebralvalley.ai/e/build-ai-teammates-hackathon)
- **When:** Sep 13 at 9:00 AM – Sep 14 at 5:00 PM (PDT)
- **Where:** Location TBA
- **Team:** [Jiangguo Zhang](https://cerebralvalley.ai/u/jiangguozhang), [Zeyu Sun](https://cerebralvalley.ai/u/zestyzapus)
- **GitHub:** https://github.com/ZeyuSun/vme50
- **Demo video:** https://youtu.be/URFiHqRyNZk
- **Gallery:** https://cerebralvalley.ai/e/build-ai-teammates-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/build-ai-teammates-hackathon/hackathon/gallery/16

VME50: Vision Means Everything in 50% Scenario

Text is not always the best way to search. Some scenes are too complex, urgent, or hard to describe in words. VME50 is an image-to-image search and interaction platform that makes vision the interface.

Upload a picture, and the system offers two options:

Where to Find It — identify the object or scene and locate it.

How to Make It — break down components into step-by-step instructions, with optional audio playback for accessibility.

Users can save results, attempt their own recreations, and upload photos back to the system. VME50 then generates a vision alignment score, which can be shared on social media for fun, competition, or collaboration.

With VME50, search becomes visual, accessible, and creative.

---

Markdown version of https://cerebralvalley.ai/e/build-ai-teammates-hackathon/hackathon/gallery/16. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt