BASALT · JOURNAL
Redacting the same area across many pages at once
A client name sits in the letterhead on all 212 pages of a report, and the same table column carries account numbers from page 40 to page 190. Drawing that rectangle 212 times is not a plan. To redact multiple pages in a PDF you mark the region once and apply it to a page range, which works because redaction is a geometric operation on a page.
Why one mark can cover many pages
A redaction mark is a rectangle in page space: an origin, a width, and a height in the coordinate system of one page. Applying it means finding every glyph whose placement falls inside that rectangle, excising those glyphs from the content stream, re-emitting the surviving text where it was, and destroying and re-encoding any image pixels inside the same box. None of that depends on what the rectangle contains, so if page 3 and page 4 share a size and a layout, the rectangle covering the header on page 3 covers the header on page 4.
That is also the one condition. A PDF can mix page sizes and rotations in a single file, and each page carries a /MediaBox, an optional /CropBox, and a /Rotate value. A landscape exhibit inserted into a portrait report has a different coordinate frame, so a rectangle expressed for the portrait pages lands somewhere unintended. Split the range wherever geometry changes.
What actually repeats in a long document
Running headers and footers are the common case. They are not one object referenced from every page, they are drawn independently into each page's content stream, which is why they behave as one occurrence per page in any search.
Letterheads and logos repeat as an image XObject placed at the same coordinates on each page. Watermarks bearing a client name often live in an optional content group so they can be toggled, and turning that layer off hides them while leaving the text fully extractable, a distinction covered in hidden layers in PDFs.
Table columns are underestimated. An appendix with an account number column in the same horizontal band across 150 pages is one rectangle applied to a range, provided the column position is stable. Check that at the first page of the range, the last, and two in the middle, because a table that reflows after a subtotal row can shift a column.
How to redact multiple pages in a PDF
Survey the document first. Note the page count, whether page sizes are uniform, and where each repeating element starts and stops.
Mark the region on a representative page, drawing the rectangle slightly larger than the visible text, because glyph bounding boxes extend past the ink and a mark that clips a descender leaves part of a character behind. Half a line height of margin is enough.
Set the page range for the mark, using the actual range rather than "all" when the element does not appear throughout, and assign the reason at the same time: privilege, PII, work product, HIPAA, or confidential. A range mark covering 200 pages is 200 entries in a log, and the reason should be attached once.
Add search based marks separately, since region marks handle furniture in a fixed place while values that move around inside body text need searching and per-hit review, covered in redacting every instance of a name.
Preview the range before applying. Stepping through the first page, the last page, and a sample in the middle with marks visible catches the rotated exhibit and the reflowed table faster than anything else.
Apply everything in one operation, written as a single generation to a new file. Applying in several passes produces an incremental update chain, and earlier revisions can still contain the unredacted page.
Then verify:
pdftotext -layout redacted.pdf - | grep -i -c "acme"
qpdf --qdf --object-streams=disable redacted.pdf expanded.pdf
grep -a -i -c "acme" expanded.pdfBoth counts should be zero. If the second is not zero while the first is, the survivors are outside the page text, in annotations, form fields, bookmark titles, or metadata. The full method is in how to check whether a PDF was really redacted.
The parts a rectangle does not reach
A region mark removes what is drawn inside that geometry on those pages, and several carriers of the same information are drawn on no page at all. The document information dictionary and the XMP packet hold author, title, producer, and often the original file path. Embedded attachments, listed with pdfdetach -list, are separate file streams, bookmark titles live in the outline tree, and form field values live in AcroForm field dictionaries even when the widget sits inside your rectangle, since painting over a widget does not clear /V.
An OCR layer is drawn on the page in rendering mode 3, so a rectangle does reach it, but only if the tool treats invisible text as text. A tool that redacts what it can see destroys the image pixels inside the box and leaves the transcription underneath. A range redaction therefore has to be paired with a document level strip in the same write. See what a PDF still carries.
Doing it more than once
Reports that reissue monthly have the same letterhead in the same place every time, and so do exports from one case management system. That makes the marking work reusable: if the geometry is stable, the sequence is stable, meaning mark this rectangle on this range, mark that column on that range, strip metadata, apply, verify. Saving that sequence and running it against next month's file removes the step where a person redraws a rectangle by hand and misses by two millimeters. The folder-wide version is in batch processing PDFs on a Mac.
Frequently asked questions
How do I redact multiple pages in a PDF at once?
Draw the mark once on a representative page and apply it to a page range, since a mark is a rectangle in page coordinates and the same coordinates apply to every page with matching geometry. Check that page sizes and rotations are uniform, preview the first, middle, and last pages, and apply all marks in one operation.
Can I apply the same redaction to every page of a PDF?
Yes, provided the pages share a coordinate frame. A rectangle is expressed relative to a page's /MediaBox and /CropBox and is affected by /Rotate, so a landscape page inserted into a portrait document will not line up with a mark drawn for portrait pages.
How do I remove a header or footer from every page of a PDF?
Mark the header band as a region on one page and apply it across the page range, so the glyphs inside that band are excised from each page's content stream rather than covered. Headers are drawn separately into every page rather than shared, so this is genuinely one removal per page.
Does redacting a watermark remove it?
Only if the redaction destroys the content rather than hiding the layer. Watermarks often sit in an optional content group, and switching that group off removes them from the rendered page while leaving the text fully extractable. Confirm afterward by decompressing the file with qpdf --qdf --object-streams=disable and grepping the expanded structure.
How do I redact the same table column across a long report?
Mark the column as a rectangle on one page and apply it to the range where the table runs, then check the column position on the first, last, and a couple of intermediate pages. Tables that reflow after subtotal rows or continuation headers can shift a column horizontally, and a mark a few points off will clip digits rather than remove them.
Should the mark be bigger than the text I am removing?
Yes, slightly. Glyph bounding boxes extend beyond the visible ink for ascenders, descenders, and side bearings, so a rectangle drawn tight to what you see can leave a fragment of a character outside the box. Roughly half a line height of margin absorbs that and any small layout drift between pages.
Is redacting page by page safer than using a page range?
No, only slower and more error prone. A range applies identical geometry to every page in it, while manual repetition introduces positioning differences and more chances to skip a page. The one thing a range needs is a check that page geometry is uniform.
Does a range redaction remove the document metadata too?
Not on its own. A region mark removes what is drawn inside a rectangle on the marked pages, while the document information dictionary, the XMP packet, embedded attachments, bookmark titles, and form field values are drawn on no page. Those need a document level strip in the same write.
Doing it in Basalt
Basalt lets you mark a region once and apply it across a page range with a reason attached, so a letterhead across two hundred pages is one decision rather than two hundred, reviewed before anything is applied. Applying excises the glyphs inside each rectangle, re-emits the surviving text in place, destroys and re-encodes image pixels in the marked area, and strips metadata, XMP, attachments, and hidden layers in the same pass, written as one generation to a new file. The original is never modified, and a verifier re-opens the written bytes with an independent parser to prove the content is gone before anything is saved. If it cannot prove that, nothing is written. Procedures save the whole sequence so next month's report runs the same way, and it is $29 once, lifetime, up to three Macs, free for 24 hours.
Redaction that proves itself
Basalt destroys the content you mark, then re-opens the file it wrote and proves the content is gone before it saves anything. Your documents never leave your Mac.