GLM-OCR

I’m the person behind GLM-OCR 👋 I started this project because I work with a lot of scanned documents and PDFs, and most OCR tools I tried were decent at plain text but fell apart once tables, formulas, or multi-column layouts were involved. The goal here isn’t to build the biggest model, but a practical one that understands document structure well enough to be useful in real workflows (papers, invoices, forms). The current model is ~0.9B params and optimized for layout + content together. The website is just a thin demo layer — everything runs through the same OCR model that’s exposed via the API. There’s a free demo, no signup required. I’m especially interested in feedback from people who process documents at scale or have strong opinions about existing OCR tools. Happy to answer any technical questions or hear what doesn’t work.