The compressor extracts every embedded raster image, re-encodes at a tunable quality (JPEG for photos, PNG for graphics), and reassembles the PDF. Text stays vector-crisp; only image content changes. Three presets: Best quality (light re-encode), Balanced (quality 80), Smallest (quality 60 + downsample to 150 DPI). File size drops of 50-80% are typical on image-heavy PDFs.
How compression works
PDFs store fonts and vector content compactly; the size bloat almost always comes from embedded images (scans, photos, screenshots). This tool decodes each image, re-encodes at your quality target, and swaps it in. Text and vector graphics are untouched — they stay perfectly sharp.
Presets and expected savings
Best quality: quality 90, no downsample — 10-25% smaller. Balanced (default): quality 80, downsample above 200 DPI — 40-65% smaller. Smallest: quality 60, downsample to 150 DPI — 60-85% smaller. Preview each preset's result before committing.
When compression won't help much
Text-only PDFs (contracts, e-books, papers) have almost no image content to compress — expect single-digit savings. Already-optimized PDFs from web export tools may have nothing left to squeeze. Vector-heavy PDFs (CAD drawings, technical illustrations) can't be compressed further with image tricks.