Documents to Markdown, parsed on your own CPU

SoMark's OXR-1.0, a 1.3B vision-language model for document parsing, compiled to WebAssembly with fusor-ai. It reads a whole page, text, tables, formulas, code or a chemical structure off an image you pick, and once loaded, no image leaves the page.

Drop, paste or choose an image

Read it as

Built with fusor-ai and fusor · model: SoMark OXR-1.0 by SoMark (OXR License) · source