A batch scan of multiple invoices, forms, or letters often comes back as one long PDF with no bookmarks and no consistent page count per document — but each one usually starts with something recognizable, like “Invoice #” or “Chapter.” That recurring text is enough to split the file back into its individual documents.

How it works

Split by Text looks for a marker phrase you specify and starts a new file wherever it appears:

  1. Upload your PDF.
  2. Type the marker word or phrase — like Chapter or Invoice # — that reliably starts each new section.
  3. Preview the resulting files.
  4. Download them individually, or as a ZIP.

Every page where that text appears (after the very first page) starts a new output file, so a stack of 10 invoices each beginning “Invoice #” splits cleanly into 10 separate PDFs.

Getting the marker text right

Matching is case-insensitive, so typing invoice will catch “Invoice #”, “INVOICE #12345”, and “invoice” alike — you don’t need to match the capitalization used in the document. What does matter is that the text is an exact substring: it can appear anywhere on the page (not just at the top), but a typo, an extra space, or a different abbreviation (“Inv.” instead of “Invoice”) won’t match. If the marker doesn’t appear anywhere in the file, you’ll simply get back the whole document as a single output, since no split points were found.

Because the match looks anywhere in a page’s extracted text rather than just the heading, pick a marker that’s unlikely to show up mid-document by coincidence. “Invoice #” is a safe choice for a batch of invoices; a bare word like “Total” is riskier, since it might also appear once or twice inside the body text of a page and trigger a split you didn’t want. If your output has more files than you expected, that’s the usual cause — try a longer, more specific phrase.

Does this work on scanned PDFs?

Only if the scan has already been through OCR. Split by Text reads the actual text layer embedded in the PDF — the same layer you could select and copy from — to find your marker. A scanned page that’s just a photo of paper, with no underlying text, has nothing for it to search, so the marker will never match and the file won’t split.

If your batch came straight off a scanner with no text layer, run it through OCR PDF first to add a searchable text layer, then bring the result back here to split it. Documents that were exported as PDF from a word processor, or scanned with OCR turned on at capture time, already have this and need no extra step.

When bookmarks are a better fit

If your document already has a proper bookmarked outline rather than a repeating text pattern, Split by Bookmarks is the more direct tool — it uses the existing outline instead of scanning for text. Use whichever matches how your document is actually structured.