Add Bates numbers to a PDF

Bates numbering puts a unique, sequential identifier on every page of a document so that a specific page can be named, cited, and found again. It is the addressing system that discovery, disclosure, and audit work runs on: once a page is SMITH-000482, it is that page in the privilege log, in the deposition transcript, and in the exhibit list, and no amount of re-sorting or re-copying changes which page anyone means.

FileGizmo stamps those numbers in your browser. Choose an unlocked PDF, set a prefix, a starting number, and a width, and the tool draws the numbers into the document on your own machine. Nothing is uploaded, there is no account, and no copy of the file reaches a server. That matters more here than on most tools: the documents people Bates-number are usually the ones they are least free to hand to a third party, and a privilege waiver is an expensive way to discover that a converter kept a copy.

Width is the whole point

The defining property of a Bates number is that every one is the same length. SMITH-000001 and SMITH-000123 align in a column, sort correctly as text, and match cleanly against a log. Take the padding away and you get SMITH-1 next to SMITH-123, which sorts 1, 10, 100, 11 and is useless as an index.

So the width is asked for rather than guessed. Choose four, six, or eight digits and every stamped page is padded to it, whatever this particular file happens to contain. A four-hundred-page volume and a two-page cover letter that belong to the same production get numbers of the same shape, which is what makes them one production rather than two.

Continuing a sequence

A production is rarely one file. Set the starting number to the next number in your sequence, keep the prefix and the width the same, and the second file continues where the first stopped. The result summary reports the first and last number it produced, so the range can be copied straight into a log or a cover letter without opening the file to check.

The prefix is usually the producing party, sometimes with a matter or volume code. A suffix is available for a custodian or a volume marker where a house style calls for one. Both are optional, and both are drawn as ordinary text, so they remain selectable and searchable in the finished document.

What the stamp does to the document

The numbers are drawn as vector text onto each selected page. The document is not rasterized, so existing text stays selectable, images are not re-encoded, and the file does not balloon. Pages that carry a rotation are stamped in the orientation a reader sees them, which is the case a naive implementation gets wrong: on a page rotated ninety degrees, the corner that looks like the bottom right is not the bottom right of the stored page box, and a number placed by arithmetic alone lands in the margin or off the page entirely.

Choose a bottom-left, bottom-center, or bottom-right position and a small, medium, or large text size. Bottom right is the conventional placement and is selected by default. Numbering follows the order of the selected pages: if the range starts at page three, page three receives the first number in the sequence.

Check before you produce

Download the result and look at it. A stamp sits at a fixed inset from the page edge, and a document whose own content runs close to that edge can end up with a number overlapping a footer or a footnote. Nothing about the numbering is destructive — the original file on your disk is untouched, and you can change the position or size and stamp it again — but the version that goes out should be one somebody has looked at.

If the document contains material that was covered rather than removed, stamp it after you have dealt with that, not before. Numbering a page does not remove anything hidden underneath a black box, and a produced set with recoverable text under a redaction is a worse problem than an unnumbered one.