Announcing An open source wireframe to app tool powered by Llama 3.2 vision. Upload a screenshot of a simple site/design & get code. 100% free and open source.

Here's the GitHub repo! Also, shoutout to @YoussefUiUx for the great design.

Tech Stack: ◆ @togethercompute's inference (AI API) ◆ @AIatMeta's Llama 3.2 Vision models ◆ @AIatMeta's Llama 3.1 405B for the LLM ◆ @codesandbox's sandpack for the sandbox ◆ @nextjs w/ tailwind & typescript ◆ @helicone_ai for AI observability ◆ @PlausibleHQ for analytics ◆ @awscloud's S3 for file uploads

How it works: I ask the Llama 3.2 vision models to describe whatever screenshot the user uploaded, then pass it to Llama 3.1 405B to actually code it. It's fairly limited in what it can handle right now – best for simple UI sketches!

Launched this as part of us at Together AI supporting the new Llama 3.2 models (including vision). Check it out!

Check out the app here!

Stats after 24h!

It would be nice to change it as OCR and extracting particular data out of pictures, invoices, packing list, delivery notes, and structure it in json or csv for handover to agent

Agreed! It has a lot of really cool use cases and I'm planning to do one with receipts potentially

You stay 👨‍🍳’ing and always be 🚢’ing! Thanks for sharing such awesome projects.

Thanks for the kind words Martin!