My tools

PDF annotate

Add text, highlights, shapes, ink and signature images to a PDF — then download a flattened copy. The file never leaves this tab.

Waiting for a PDF

Annotate a PDF without uploading it.

Drop a PDF on this card, paste one with ⌘V / Ctrl+V, or choose a file. It opens right here — there is no Upload button.

One PDF at a time · up to 50 MB · up to 2,000 pages · images up to 10 MB (PNG, JPEG or WebP).

The PDF stays in this tab: nothing is uploaded, stored or logged, and the page sends no network request.

How to use it

  1. Drop a PDF on the card, paste one with ⌘V / Ctrl+V, or press “Choose PDF”. The first page is drawn right away — there is no Upload button and nothing leaves this tab.
  2. Pick a tool: Text opens a box where you click, Highlight and Rectangle are dragged out, Ink follows the pointer, and Image places a PNG, JPEG or WebP you choose (a signature photo works). Use Select to move a mark or Delete selected to remove it. Undo and Redo cover every change on every page.
  3. Move between pages with the arrows or the page number. When the marks look right, press “Download annotated PDF”: the tool draws each marked page at 2× and writes those pixels into a new file. The status bar shows the output size when it is done.

Example

report.pdf — 6 pages

Page 2 · highlight the total
Page 3 · place a signature image on the line
Page 4 · add a text note

report-annotated.pdf

report-annotated.pdf
All marks are flattened into the pages
Source file unchanged

The downloaded PDF contains drawn pixels, not editable annotation objects — that is what makes it look the same everywhere. Arabic and Chinese notes are shaped with the fonts installed on your device.

Rules, limits and known issues

  • Nothing is uploaded: the PDF is read with the File API and annotated in this tab. The page sends no network request, writes nothing to storage and keeps no history — a reload starts empty.
  • Marks are flattened on save. The new PDF cannot edit or remove them again, and the source file is never modified. This is not a full PDF editor: existing text, images, annotations and page order are not changed.
  • Highlight is a translucent rectangle, not a text-aware highlight. It does not snap to words or lines, so draw it over the text you mean.
  • Text marks become pixels: they cannot be selected, copied or searched in the saved PDF. Appearance depends on the fonts installed on your device, and there is no font upload.
  • Signature means an image (PNG, JPEG or WebP) placed on the page. It is not a cryptographic or certificate signature, and rebuilding the file invalidates any existing digital signature.
  • One PDF at a time, 50 MB, 2,000 pages; one image at a time, 10 MB. SVG images are rejected because they can contain scripts.
  • Password-protected PDFs are rejected — remove the password and save a fresh copy first. Severely damaged PDFs are not repaired.
  • Rotated pages and CropBox offsets are supported; the mark is placed where you saw it on screen. Pages whose long side would exceed 4,096 px in the annotation layer are scaled down for export.
  • Marks live in memory only. There is no autosave, no history and no recovery after a reload or tab close.
  • The output name replaces path separators, control characters and reserved characters, trims the ends, cuts long names and appends “-annotated.pdf”.
  • Existing bookmarks, forms and annotations are left in place, but the file is rewritten, so PDF/A and PDF/X conformance is not guaranteed.
  • Rendering uses Mozilla’s PDF.js and writing uses pdf-lib, both inside this tab. A few unusual PDFs (damaged fonts, exotic colour spaces) can render incompletely.

Other document tools that also run entirely inside this tab.

PDF page numbers

PDF text extract

PDF metadata