PDF compression that keeps the text and fonts

An online tool for compressing PDFs. Instead of simply re-saving the file, it takes the document apart and finds where the bytes actually live: images, fonts, page streams or metadata. Only images whose resolution is excessive for the size they are actually drawn at get recompressed — text, fonts, vectors, links and the outline stay untouched. There is a fit-under-25-MB mode and a before/after page comparison right in the browser.

How to compress a PDF

Upload the PDF

Drop the file or pick it from your device. Reading it takes a fraction of a second.

Look at what is inside

The tool shows where the weight is and proposes settings suited to this kind of document.

Auto, a target size, or your own DPI

Auto, fit a size, lossless, or your own DPI and quality.

Check a page before and after

Compare the page before and after, confirm the text is still there, and download the result.

Cut the file size without losing text, fonts or bookmarks

Drop a document here

PDF ·

Loading file...

What the tool does

Shows what the file is made of

Before you press anything, you see where the weight sits: images, fonts, page graphics, metadata. It is immediately clear whether there is anything to gain, and why.

Recompresses images, not pages

Each image is judged by how large it actually appears on the page. An image already within budget is not touched at all — re-encoding it would only degrade it.

Text stays text

Search, selection and copying keep working. Fonts, links, the outline and page labels survive — rasterization is opt-in only, and it says plainly what it costs.

Fit-a-size mode

Ask for 25 MB for Gmail or 10 MB for an upload form, and the tool finds the quality itself, keeping the best result that still fits.

Vector PDF optimization

Documents where text has been converted to curves — forms, drawings, anything printed to PDF — shrink through coordinate precision and stream repacking, with no visible difference.

Checks its own work

It reports how many pages survived, whether the text layer is intact, and lets you wipe between the page before and after — you see the result before the file goes anywhere.

Typical uses

Get a 50 MB scanned contract under the 25 MB Gmail attachment limit.
Shrink a photo-heavy deck to fit a portal that caps uploads at 10 MB.
Compress a stack of scans for archiving while keeping small print legible.
Compress a 200-page report with its outline, links and text search all still working.
Squeeze a form whose text was converted to curves, where an ordinary re-save gains nothing.

How this differs

Most online compressors offer a row of presets and turn the pages into pictures. The file does get smaller — but text search, selection, links and the outline go with the weight. This one works differently: the document is taken apart, and only what is genuinely oversized gets compressed. Each image is judged against the size it is actually drawn at, images already optimal are left alone, and if the result is not smaller than the input, you get the input back rather than a proud "−0%". Afterwards the tool checks that the pages are all there and the text layer is intact, and shows you the page before and after.

Frequently asked questions

Will the text still be searchable?

Yes. In every mode except manual rasterization, the text layer, fonts and links are preserved. After compressing, the tool tells you whether the text survived.

How much smaller will the file get?

It depends on what the file is made of, and you can see that on the composition card before compressing. Scans and photo-heavy documents typically lose 60–95%, vector forms around 25%, and a font-only text document 10–25%.

What if the file is already compressed?

The tool notices and hands back the original unchanged, saying plainly that there was nothing to gain. The result can never come out larger than the input.

How is lossless mode different?

It never touches an image or a coordinate: it only repacks streams, merges identical objects and drops anything nothing references. The rendering stays identical.

Are bookmarks and the outline preserved?

Yes, in every mode except rasterization — that one rebuilds the document from pictures and cannot carry the outline across.

Why were some images left as they were?

Either they are already compressed enough, or they use JPEG 2000, CCITT G4 or JBIG2, which a browser cannot re-encode. The tool says so on the composition card.

💡 Want us to improve this tool just for you?

We can! Just send us a quick message with your idea. If you'd like to discuss it in detail, leave your email and we'll get back to you. You can stay anonymous.

How do you rate this tool?

Thank you for your rating!

Change my rating

Want to leave a comment?

Want to share more? Leave a comment!

Thank you! Your comment will appear after moderation.

Who is this tool for?

Published 21 May 2026 Author: The Supreme