Compress PDF
Shrink PDFs by re-encoding embedded JPEG images at lower quality. Best for scans, decks, and photo-heavy reports. Runs in your browser.
Runs in your browser — nothing uploaded
Shrink a PDF by re-encoding its embedded JPEG images at lower quality. Best results on scan-heavy files, photo-heavy decks, and screenshot-laden reports — the documents that are big in the first place.
Private by design — your data never leaves your device
How to use it
No account, no upload — it all happens on your device.
1
Drop a PDF onto the box, or click to choose one.
2
Pick a quality preset: high keeps photos sharp, low produces the smallest file.
3
Click Compress PDF. If there are savings, a download button appears.
What this compresses (and what it doesn't)
Knowing the lever helps predict the result.
- Re-encodes embedded JPEGs at the chosen quality. This is where the savings come from.
- Keeps text and vectors untouched. Selectable text stays selectable. Charts stay sharp.
- Skips non-JPEG images and chained filters. Lossless image formats and unusual filter chains are left alone — re-encoding them safely is more work than v1 takes on.
- Doesn't downsample resolution. A 300 DPI image stays 300 DPI; only its JPEG quality changes. A resolution-aware version is in the backlog.
When compression won't help
- Pure-text PDFs.Text and vector data is already compact; there's no fat to trim.
- Already-compressed PDFs. Running a small PDF through a compressor a second time produces marginal or no savings.
- PDFs from designers who flattened images to vectors. The bytes are in the vector data, not in JPEG streams.
For the deep version of this story, see the blog post on how PDF compression actually works — it covers DPI, fonts, structural compression, and where the real bytes live in different kinds of PDFs.
Frequently asked
How much smaller will my PDF get?
It depends on what's inside. Scan-heavy PDFs and photo-heavy decks often shrink 40–80% at medium quality. Text-only documents and PDFs that have already been compressed see little change. The result panel shows the actual saving so you can decide whether to keep it.
Will the text get fuzzy?
No. We only re-encode embedded JPEG images. Text and vector graphics in the PDF stay untouched. The visible quality drop is in photos and scans, and it's tuned by the quality setting.
Is the PDF uploaded?
No. The PDF is parsed, image streams are re-encoded with the browser's canvas API, and the file is rewritten — all inside your browser tab. Nothing leaves your device.