PDFs loaded in browser should automatically be OCR'd to be searchable. This can be done locally using various FOSS OCR engines. This is a feature I'd be happy to build if Orion was open source.
Users often open scanned/image-based PDFs where text can’t be searched, copied, or read by accessibility tools. A built-in OCR feature would:
Enable search: Cmd+F works inside scanned papers, contracts, receipts, etc.
Allow copying/quoting: Users can select and copy selections from text instead of retyping.
Improve accessibility: Screen readers can access text in image-only PDFs.
Integrate smoothly: Orion’s PDF viewer could auto-detect image-only PDFs and offer “Recognize text,” or add a toggle to overlay selectable text.
Other implementations: Chrome’s PDF OCR, macOS Preview (Live Text), and Adobe Acrobat show how valuable this is.
Benefit to Orion: Keeps text recognition local/private, avoids third-party tools, and makes Orion more research- and user-friendly by turning static scans into fully usable PDFs.
