The whole page, exported as one PDF.

When PDF is the right format

A PNG is easy to paste into a chat or a ticket. A PDF is what most legal, support and compliance workflows expect as a record: the terms of service as they read today, the exact state of a client’s page on a specific date, or evidence attached to a support case. Naoya’s PDF export exists for that second job.

Nothing about the capture itself changes between formats — the same scroll, tile and stitch pipeline runs either way. The format choice only affects how the final image is packaged for saving.

Three scenarios show up over and over: proving a landing page changed by comparing before and after, documenting a bug that only appears 4,000 pixels down a page, or filing away a terms-of-service page exactly as it read on the day it was published. All three need the page as it actually was, not a screenshot cropped to what fit on screen — which is what a full-page PDF is for.

How the split actually works

Browsers cap how tall a single canvas can be. A page long enough to exceed that limit cannot be held as one image no matter how much memory is available, so a capture past that limit is cut into slices and laid out one per page instead.

The cut is a simple loop: every slice except the last is exactly one page’s worth of the image, top to bottom; the last slice takes whatever remains. That is why a page never ends with a band of blank space at the bottom, and content is never cropped short at the end — the final slice is sized to its own remainder, not forced to the same height as the others. Each page is sized in points, the PDF unit, using one of three standard paper sizes — A4, Letter or Legal — in either portrait or landscape orientation; the image itself is untouched, only where each page’s boundary falls changes.

Same capture, three file formats

PNG, JPEG and PDF all come from the same stitched bitmap; the difference is only in how that bitmap is packaged, not in what was captured. PNG and JPEG re-encode it at its native captured size, pixel for pixel. PDF is the one format that resizes anything, and only because a page needs physical dimensions: the image is scaled so its width exactly fills the paper’s width in points, and its height follows from that same scale factor — which is also what decides where each page boundary falls in the split above. JPEG’s own quality is clamped to the range the Canvas encoder actually accepts, 0 to 1; left unset, it falls back to 0.92, which is not a number Naoya picked but the Canvas API’s own default for toBlob/convertToBlob.

What PDF export does not do

It does not add a signature, a watermark or a redaction — the export is the page as captured, nothing drawn on top of it. If you need to annotate or redact a capture, that is outside Naoya’s current scope; the product is judged on capture fidelity first.