BASALT · JOURNAL

An Acrobat alternative for Mac when redaction is the job

2026-08-06 · adobe acrobat alternative mac

Most people looking for an Adobe Acrobat alternative on Mac want to stop paying a subscription for a program they open four times a month. If your four times a month include redacting documents before they go to opposing counsel, a regulator, or a public records requester, the calculus is different. You are not shopping for a PDF viewer. You are shopping for a tool whose failure mode is a disclosure.

This article is about what redaction actually requires from a PDF application, and how to evaluate any Acrobat replacement against that standard rather than against a feature checklist.

What redaction is, at the file level

A PDF page is a content stream: a sequence of operators that a renderer executes to paint marks. Text is drawn by text-showing operators, Tj, TJ, ', and ", which take strings of character codes and map them through a font's encoding to glyphs. The visible page is the output of running that program. The file is the program.

Redaction means editing the program so the sensitive operands are gone. If a name appears in a TJ array, the character codes for that name have to be removed from the array and the surrounding text has to keep its position, which usually means splitting the run and re-emitting the survivors with corrected offsets. Anything short of that leaves the codes in the file where any parser can read them.

Drawing a black rectangle does not edit the program. It appends a fill operator that paints over the earlier marks. The text operators are still there, still executed, still extractable with a copy and paste or a command line tool. This is the single most common redaction failure and it is discussed in more detail in the black box myth.

Where Acrobat sets the bar, and where the bar actually is

Acrobat's redaction tool does perform real content removal, and it is fair to say it is the reference implementation most legal teams have used. The practical complaints are elsewhere: the subscription, the size of the install, the amount of the application that is not redaction, and the fact that the redaction and the sanitization of hidden data are separate commands a user has to remember to run in the right order.

That last point is the interesting one. Removing text from a page does not remove:

• Document information dictionary fields and the XMP metadata packet, which frequently carry author names, original filenames, and revision history.
• Embedded file attachments.
• Annotation contents, including comment text, sticky notes, and the /Contents value of markup annotations that may quote the passage they annotate.
• Form field values in the AcroForm dictionary, which are stored independently of the page's visible appearance stream.
• Optional content groups, meaning layers that are simply switched off rather than absent.
• Stale page thumbnails and cached appearance streams generated before the edit.
• Prior document states retained through the incremental update chain, where a PDF is saved by appending a new cross-reference section and leaving the old objects intact earlier in the file.

That last mechanism deserves emphasis. An incremental save does not rewrite the document. It appends. The pre-redaction version of an object can remain in the bytes ahead of the new one, reachable by anyone who parses the file rather than viewing it. A tool that redacts correctly and then saves incrementally has produced a file that leaks. Hidden data in PDFs covers each of these surfaces in detail.

So the bar for an Acrobat alternative is not "has a redaction button." It is: does it remove the operands, does it remove every parallel copy of the same information, and does it write a single clean generation rather than an append.

Free and built-in options, briefly

Preview on macOS is on every Mac and is the tool most people reach for first. Its markup tools draw. It does not excise text from content streams, and its redaction-adjacent features have historically produced files where the underlying text remains selectable. Why Preview is not safe for redaction walks through the specifics.

Browser print-to-PDF is sometimes suggested as a flattening trick. It does destroy the text layer, which is genuinely useful, but it also destroys everything else: searchability, Bates numbers, bookmarks, and fidelity. It rasterizes at the printer's resolution, and it gives you no record of what was removed. It is a blunt instrument that occasionally does the right thing for the wrong reason.

Browser-based PDF editors solve the convenience problem by uploading the document to a server. For regulated material that is a transfer to a third party, and it is a transfer you have to account for. If the document contains personal data or protected health information, the upload itself is the compliance event, before anyone has clicked redact. See the risk in cloud PDF tools.

The question nobody asks: how do you know it worked

Every redaction tool tells you it succeeded. The interesting engineering question is what that claim is based on. In most implementations it means the code path completed without throwing an error. The tool is asserting its own correctness using the same data structures it just wrote.

The stronger design is to re-open the finished file with an independent parser, one that does not share state with the writer, and search the result for the content that was supposed to be gone. Extract all text. Check every annotation value, every form field, every attachment, every metadata packet. If the target strings appear anywhere, the write failed and the file should never reach disk.

This is a meaningful difference in posture. It converts "we removed it" into "we removed it and then proved the removal against the bytes we actually wrote." How to verify a PDF redaction describes how to run that check yourself, independent of any vendor.

Evaluating an alternative

A workable checklist when you are testing any Acrobat replacement on a sample document:

1. Redact a word in the middle of a paragraph. Then open the output and select the paragraph. Copy it. Paste it into a plain text editor. The word must be absent and the surrounding words must retain their spacing.
2. Run a command line text extractor over the output. Anything that survives extraction survives discovery.
3. Check the file size and structure. If the output is the input plus a few kilobytes, you probably got an incremental update rather than a full rewrite.
4. Inspect the metadata. Original filename, author, and producer fields have all leaked identities in real matters.
5. Redact a region of a scanned page and then extract text again. If the document has an OCR layer, the invisible text under the mark is the leak. See redacting scanned documents.
6. Redact the same file twice with identical inputs and compare checksums. Deterministic output tells you the tool is doing something defined rather than something incidental.