Frequently Asked Questions

Anything PDFKit can open and qpdf can parse: Latin, CJK, and Arabic text, RTL scripts, embedded fonts, subsetted fonts, Type0 composite fonts, and diacritics. Tested up to 100 pages and 200 MB file size. Encrypted PDFs need to be unlocked first — Redaktor refuses to write an output for a document it cannot safely re-parse.

Interactive form fields store their values in the /V entry of the field’s widget annotation, which lives outside the page content stream. Redaktor flattens the form before redaction: PDFKit bakes the widget appearances into the page content stream, and the engine then removes the glyphs inside your boxes along with the rest of the content. In the normal case there is no warning — the form fields are gone from the output. If the automatic flatten fails on an unusually structured AcroForm, the export report warns that form field content may not be fully redacted and asks you to flatten the form yourself in Preview or Acrobat before trying again.

The text-extraction channel reported clean, so no text scraper will find anything. But Apple Vision OCR detected readable characters in the rendered pixels of a redaction region. Usually this means the PDF contains a scanned image or a raster overlay with baked-in text. The default surgical engine works on the content stream, not on pixel data, so it cannot remove characters from inside an image. Use Flatten as image PDF when this matters: Redaktor renders each page to pixels, burns your redaction boxes into those pixels, and writes a new image-only PDF without source text layers, forms, annotations, or source metadata. Tradeoff: text selection, search, links, forms, and annotations are removed, and the file can become much larger.

Outlined text is stored in the PDF as vector paths (lines and curves that happen to look like letters) instead of glyph operators. There is no Tj / TJ in the content stream to remove. The engine correctly reports zero modifications, and the OCR channel flags the region. Flatten as image PDF is the fallback for this case, with the same tradeoffs: no text selection, no search, and no interactive PDF elements in the export.

Typical documents (10 to 50 pages) export in seconds. The surgical engine only rewrites the affected content streams; the rest of the PDF passes through untouched. Flatten as image PDF takes longer because it rerenders every page to pixels. A progress bar is shown during export, and you can cancel at any time.

Practically no. The entire redaction pipeline runs offline: PDFKit load, qpdf content-stream rewrite, Apple Vision OCR on rendered regions, file write through NSSavePanel. The app makes network calls only outside the document workflow:

1. Apple’s StoreKit for the optional Tip Jar.
2. Apple’s SKStoreReviewController for the in-app review prompt.
3. An anonymous GET to cagatay.io/apps/catalog.json when you open the Settings window, to populate the More Apps cross-promo list. Response is cached locally.

None of these paths transmit personal data or document content.

Only a small set of app preferences in UserDefaults inside the macOS sandbox: a successful-export counter and last-shown date for the review-prompt gating, a flag for whether the support-mail dialog includes system info, six flags tracking first-use ambient guidance, and the cached More Apps catalog JSON (capped at 100 KB). No PDF contents, no document paths beyond what macOS keeps in the system Recents list, no analytics, no browser cache.

No. Your PDFs exist solely in your Mac’s memory and are discarded when you close them. There is no cloud backend, no automatic caching, no iCloud sync. The full privacy policy is at cagatay.io/apps/redaktor/privacy.

No. One-time purchase, yours forever. All features are unlocked from the start. The optional Tip Jar in Settings is purely voluntary and does not affect functionality.

Yes, free via the Mac App Store, delivered automatically as long as the app is installed.

Redaktor requires macOS 14 Sonoma or later. Tested on both Apple Silicon and Intel Macs.

Common causes:

• No write access at the chosen location. Pick a different folder.
• Attempting to overwrite the source. Redaktor blocks this; choose a new filename.
• PDF damaged. Open it in Apple Preview and re-export as a new PDF.
• Encrypted PDF. Unlock it first (Preview → Export As, without the password).

If the problem persists, email support@cagatay.io, ideally with the version number (Settings → About).

Either the text-extraction channel found readable text inside a redacted region, or Apple Vision OCR did. The receipt lists which pages and what residue. Most common causes: the redaction box did not fully cover the sensitive text — widen it by a few points and re-export; or the content was a scanned image or outlined text — use Flatten as image PDF from the receipt to redo the export by burning the boxes into pixels.

PDF glyphs can extend a sub-point past the visible selection at the edges (descenders on "g"/"y", italic tails). Redaktor pads every converted rect by 1 pt on each side to catch this. If you still see residue, widen the box by a few more points before exporting.