iOS 18.3 does not fix this bug, either! Apple, please!!, fix this bug for iOS 18.4.

Is this iOS 18 only, or does it happen on iOS 17.x and earlier, too?

Hi Randy, seems you are working with CoreGraphics PDF lib. As iOS 18.x still has a PDFKit bug (see here: https://developer.apple.com/forums/thread/762788 ) I have been looking into CG myself. Getting text with a CG Scanner out of a PDF is quite doable, but getting the bounding boxes of those texts is quite a challenge. Since you seem to have experience with Core Graphics PDF routines, can you steer me to same sample code for getting bounding boxes of characters/words/text blocks out of a PDF page with Core Graphics? Thanks for any help! Regards, Klaus.

iOS 18.4 beta 1 and 2 do not fix the issues. Selection.bounds(at: page) delivers more correct results than page.characterBounds(at: index). But not as good as characterBounds(at) for iOS 17 and earlier. // page: PDFPage for (index, character) in page.enumerated() { if character == "\n" || character == "\r" { continue } let range = NSRange(location: index, length: 1) guard let selection = page.selection(for: range) else { continue } let charBounds = selection.bounds(for: page) } Do we still have a chance for a PDFKit bug fix from Apple? Does Apple really think that PDFs are dead and it's not worth keeping its PDF libraries up to date? What about the billions of PDF documents out there and the millions more generated every day? If you need character and word accurate text extraction with precise layout data, LLMs are still no substitute.

iOS 18.4 beta 4 does not fix the issue.

iOS 18.5 still has the same issue.

Any news from Apple on the issue?

You can disable the visibility of the Format panel with .textInputFormattingControlVisibility(.hidden, for: .all)