I am looking for a computer science student or someone interested who wants to help with the next iteration of Screenshot to Layout.

Here’s a demo of Screenshot to Layout 2.0 and how I am working on it. I share a bit of the history and vision of the project.

Since this project is so experimental, it’s generally something that I’m doing hacking away late nights, but I’m learning a ton. And I think you would too 🤓

• We have our own MCP server with tool calls forked over from Figsor
• We use Claude skills and are inspired by Figma’s own skills to make things better as well as the figma-console-mcp project
• There’s a raw_compute branch where we use pixel level analysis, which we’re now bringing to the main branch
• We’re interested in an agent that can improve itself via a feedback loop (this is hot in the AI community right now

FigOCR as a project live, you can find info via FigOCR.com (I might rename this back to Screenshot to Layout eventually if we can hit the quality level we want).

If you are interested to work together, get in touch!