Local system onlineAPI docs
LESS FRICTION. MORE FORWARD.

Free PDF compressor under 1 MB

From “file too large” to ready to go. Pick a size or a destination. We’ll handle the rest.

1

Add your file

Up to — MB on server
+

Drop it here. We’ll make it fit.

Drag & drop your file, or browse from your device

Images, PDFs, documents, video & audio
ImagesPDFs & docsVideoAudio
2

Tell us where it needs to fit

Popular sizes
Private by defaultRust/WASM in your browser → Rust edge → server only if needed.
Quality preserved where possible. Every output validated.
Checked, not guessedDeterministic validation on every output
Only yours to accessPrivate storage and short-lived download links
Here for a little whileDelete now, or let automatic cleanup do it

How to use this target

A mostly text-based PDF is very different from a scanned document. Text and vector content may already be compact, while scanned pages store large images. This target uses 1,000,000 bytes rather than assuming the uploader means 1 MiB.

  1. Start with the original PDF. Check whether you can select text, or whether every page is a scanned picture.
  2. Choose the 1 MB preset and enable searchable text only if that is required. Browser-only mode can perform structural cleanup but does not rebuild scans with OCR.
  3. Inspect every output page, verify page count and test text selection or search. Download only after the output passes the technical checks.

What to check before uploading

Not every PDF can fit 1 MB without losing necessary information. Encrypted, signed, form-heavy or active-content files may be rejected to avoid unsafe or destructive changes. Large scanned PDFs may require temporary server processing; choose on-device-only to refuse that transfer.

Common questions

Does PDF compression remove pages?

This workflow preserves page count rather than deleting pages to meet a target. If all required pages cannot fit within the allowed settings, use another permitted approach.

Will searchable text remain searchable?

Existing native text is preserved by the supported cleanup path. Scans may need OCR, whose output still needs review for names, numbers and small print.

Other targets