PDF

Remove handwriting from a PDF

Upload a PDF and each page is analyzed on its own — handwriting is erased where it exists, and pages that are already clean are left completely untouched.

Drop a PDF or image here

or browse your files — PDF, JPG, PNG or WEBP, up to 25MB

Your document is processed entirely in your browser. It is never uploaded.

Why page-by-page matters

A lot of PDFs mix clean, typed pages with a handful of pages someone wrote on — a signed form buried in a clean packet, or a scanned worksheet with answers on only some pages. Reprocessing every page through an image pipeline would throw away the sharp, selectable text on the clean ones for no reason. Instead, each page is rendered and checked for handwriting independently: pages with none are copied into the result as the original PDF page object (vector text and all); pages with handwriting are rebuilt from a cleaned image.

Two kinds of "handwriting" in a PDF

There's handwriting that's part of a scanned or photographed page — physically written with a pen before the page was digitized — and there's handwriting added digitally inside a PDF editor, stored as a separate annotation object (for example, a stylus drawing made in Preview, GoodNotes, or Adobe's Draw tool). Both are handled: digital ink annotations are removed directly and losslessly from the PDF's own data before anything else happens, and scanned/photographed handwriting goes through the AI detection and inpainting pipeline described on the how it works page.

What stays exactly the same

  • Page order, count, and dimensions
  • Printed text, equations, tables, and diagrams outside handwriting regions
  • Pages with no handwriting at all, copied as vector data with no quality loss
Working with a signed contract or an official record? Use this to remove your own draft markup — not to alter a document's legal or evidentiary content. See removing pen marks from a PDF for more on that distinction.

FAQ

Questions about this

Will this change pages that have no handwriting?

No. Each page is checked independently. A page with no handwriting is copied into the output PDF exactly as it was in the source file — including its original vector text, if the PDF has any — instead of being rasterized and rebuilt.

How many pages can I upload?

Up to 20 pages per PDF in the current configuration, to keep processing time and cost predictable. Larger documents can be split into parts.

Does it handle PDF annotations, not just scanned handwriting?

Yes, as a separate lossless step: native PDF 'Ink' annotations (freehand strokes added in a PDF editor) are stripped directly from the PDF's annotation data, before any image-based detection runs.

Will the page size or orientation change?

No. Each output page matches the source page's dimensions and orientation. Nothing is cropped, resized, or unexpectedly rotated.