alam00000/bentopdf v2.8.10

on GitHub

Thanks to @marek-mraz, and @musicsms for responsibly disclosing these.

A little update from me

Hey everyone!

First off, I want to apologize for the slower releases lately. I haven't been doing too well health wise, so things slowed down quite a bit, and I haven't been able to review PRs or reply to issues as fast as possible

I'm doing much better now, though, and I'm happy to say that BentoPDF is going back to weekly releases!

I've been working on quite a few things behind the scenes, and the next update is going to be a big one. Here's what's coming:

• Measurement tools: You'll be able to measure distances, perimeters, and areas directly on your PDFs.
• Favorite tools: Pin the tools you use most so you don't have to keep looking for them.
• Air-gapped by default: I'm gradually removing dependencies on third-party engines like Ghostscript and PyMuPDF and replacing them with our own engines, [Kura](https://www.bentopdf.com/kura) and [Hyper Compress](https://www.bentopdf.com/hyper-compress).
• A much faster Organize: I've spent a lot of time optimizing the Organize tool, and the performance improvements are pretty significant.
• Digital signatures: PAdES signatures with timestamps and long-term validation, along with certification signatures.
• Dark mode and color-blind modes

Most of my time, though, went into building a new lightweight, SOTA OCR engine for PDFs. It's fast, it performs at the level of vision language models (VLMs). Beyond reading text, it handles:

• Tables: Recognizes tables, including merged cells, and reconstructs them into proper rows and columns that you can export to Excel or HTML.
• Formulas: Recognizes mathematical expressions and converts them into LaTeX.
• Markdown export: Converts entire pages into structured Markdown, preserving headings, lists, tables, and formulas.
• JSON export: Provides recognized text along with its position on the page and confidence scores.
• Layout detection: Identifies paragraphs, headings, lists, tables, captions, footnotes, forms, and more.
• Reading order: Handles multi-column documents, magazines, and complicated layouts without mixing everything up.
• PDF tagging: Can help turn unstructured PDFs into properly tagged documents.

And yes, at one point I genuinely thought training an OCR model on my poor GTX 1650 was a good idea.

Thank you for using BentoPDF and the support. More updates are coming soon ❤️

What's Changed

• build(deps-dev): bump handlebars from 4.7.9 to 4.7.10 by @dependabot[bot] in #926
• build(deps): bump source-map-js from 1.2.1 to 1.2.2 by @dependabot[bot] in #927
• build(deps): bump PDF engine and viewer to 051abe5b72a9 by @github-actions[bot] in #932

Full Changelog: v2.8.9...v2.8.10