i
Open a heavy PDF or an oversized photograph and choose how hard to squeeze it. It is re-encoded in this tab, the before and after sizes are measured, and nothing is uploaded.
1. Open a file
Drop a PDF or a picture here, or click to choosePDF, JPEG, PNG, WebP or BMP. Nothing is uploaded; it is opened inside this tab
2. How hard to squeeze
.
Leave this on unless the PDF is a pure scan with no real text anyway.
A heavy file in, a lighter one out. A PDF or a picture; the
same three settings either way.
The usual reason a PDF is enormous is that it is a scan: big photographs
of pages at a far higher resolution than anyone needs. This re-renders each page
at a sensible size and re-compresses it, which on a scan can cut the file by ten
times. The text is laid back over each page invisibly, so you can still select
and search it. The invisible layer is a Latin-only font, so Latin text stays
selectable while other alphabets become part of the picture.
The usual reason a photograph is enormous is the same thing wearing a different
hat: a 12 megapixel picture of a receipt. A picture is scaled down to a sensible
longest side and re-encoded, which is the whole of it.
On a PDF that is already mostly crisp vector text, or a picture that is
already small, this will help little or may even make it larger, because there is
nothing oversized to shrink. So it shows you the before and after and lets you
keep the original if it won. It never overwrites anything; you download a new file.
If you want to choose the exact dimensions, the format and the quality yourself,
that is image tools. This screen is the other kind of
tool: you say smaller, it decides.
Is this actually private?
Nothing is uploaded. Static page, no endpoint, no account. The file is rendered and rewritten in your browser.
A picture is scaled and re-encoded, and it says which. Light, Medium and Strong are a longest side and a JPEG quality: 2600px, 1800px and 1200px. A picture already smaller than that is not enlarged to meet it. Something with transparency comes back as WebP so the transparency survives, or as PNG where the browser cannot write WebP; anything opaque comes back as JPEG, which is what makes it small.
An animated GIF is turned away rather than flattened. Re-encoding one here would hand back a single frame, which is not a smaller version of what you gave it. Video tools is where a GIF belongs.
It re-renders; it does not optimise in place. Every page becomes a re-compressed image with an invisible text layer over it. That is what shrinks a scan dramatically, and what makes it the wrong tool for a document that was already efficient vector text, where it will show no gain and you keep the original.
Nothing is thrown away silently. It reports the before and after size, and if the result is not smaller it says so. You choose which to download.
Fonts and vector art become pixels. A shrunk page is an image, so zooming in far enough shows softness a vector PDF would not. At Light strength on a normal document that is hard to notice; at Strong on a dense page it is visible.
Big documents use memory and time. Each page is rendered at full size before being compressed, so a 300-page scan is real work for a phone. It runs a page at a time so the tab does not freeze.
Nothing here is for sale, nothing is tracked, and there is no account to make.
If something saved you an hour: 83TQcTwusSQ4WKbPQE5osrF3cR4GWe2zmcNWeozK6BSqHSaeLvjUVe476ouVwLKn1uVwEFcbJQvnme7W6dTV5SB93x45DEy
You can
check that address here before sending anything.