GLM-OCR reads text from photos, scans and PDF files. You upload a document by dragging it onto the page, and the tool returns editable text. It goes beyond plain text by recognising tables, which it can output as Markdown, and mathematical formulas, which it can turn into LaTeX.
The underlying model is fairly small, at 0.9 billion parameters, and the site mentions support for more than eight languages, an API for developers and a GitHub page. It works in the browser, so there is nothing to install for basic use.
