Every page produced in litigation gets a name it will keep for the life of the case: ACME-000417. When a deposition refers to it, a brief cites it, or a judge asks for it, that label is how everyone finds the same page. That is Bates numbering, and it has rules that ordinary page numbering does not.
The rules
- Unique. No two pages in a production share a number. Ever.
- Sequential and continuous. Numbers increase by one, page by page, through every document in the production — and through every batch. The second production picks up exactly where the first stopped.
- Prefixed. A short code identifies the producing party or the matter: ACME, DEF, SMITH-. Agreed with the other side before anything is stamped.
- Zero-padded. ACME-000417, not ACME-417, so numbers sort correctly as text and the format never changes as the count grows. Six digits is conventional.
- On every page. Including cover sheets, blanks and the backs of two-sided scans if they were produced.
- Legible and out of the way. Usually the bottom-right corner, small, in a margin the content does not use.
Break continuity — restart at 1 for a new batch, skip a range, produce a duplicate — and the production is challenged. The rules exist because the labels are evidence.
Doing it on your own machine
Discovery documents are, by definition, the ones with the highest stakes for confidentiality: privileged communications, personal data, trade secrets. Uploading them to a stamping website means a third party holds an unredacted copy of a production, however briefly.
Header, footer & Bates numbers applies the stamps inside your browser. The document is read into memory, the labels are drawn into each page’s content, and the result is saved locally — nothing is transmitted.
What the tool gives you
- Tokens for the running number and the page number, so a footer can read ACME-{bates} and a header Page {page} of {pages} at the same time.
- A start value, which is how continuity across batches works.
- Prefix and padding to match the agreed format.
- Page ranges, so a privilege log or a cover sheet that must not be stamped can be excluded, and so a stamp never lands on an excluded page by accident.
- Position and size for the running heads and the label.
Two constraints worth knowing: the stamp is set in a standard font whose encoding covers Latin text, so prefixes should use plain letters and digits; and the stamp is placed relative to the visible page area, so on a scan whose page box starts off the corner the margin is measured from where the page actually shows.
Step by step for a scanned production
- Prepare the scans. Auto crop removes scanner borders so the stamp lands in a clean margin.
- OCR with OCR PDF if the production must be searchable — most protocols require it.
- Open Header, footer & Bates numbers. Set the prefix, the padding width, and the start value (1 for the first batch; the previous last number plus one for any later batch).
- Choose the position — bottom right is conventional — and a page range if any pages are excluded.
- Preview the first, a middle and the last page. Confirm the number sequence reads correctly and does not overlap content.
- Save. Record the range — first and last number, and the batch — in your production log.
- Protect PDF if the production is delivered encrypted, or Flatten first if it carries annotations.
The Prepare a scanned contract for filing workflow runs this sequence — trim, recognise, number, protect — passing the file between tools without re-selecting it.
Checks before you produce
- First page, last page: does the number match the log?
- Any page where the stamp collides with content? Move it, do not shrink it below legibility.
- Excluded pages actually unstamped?
- Total pages equals last number minus first number plus one? If not, something was skipped or doubled.