PDF Compressor uses pdf-lib to rebuild a PDF around object streams and a cross-reference stream. These pack the document's internal objects into compressed streams instead of listing them in a plain cross-reference table, which is where the saving comes from. Files written by older software often use the plain table and can lose roughly half their size; a file already built this way is returned unchanged rather than rebuilt. Image content is not re-encoded, so image-heavy PDFs see little movement either way.
This tool rebuilds the PDF using object streams and a cross-reference stream, which pack the document's internal objects more tightly than a classic cross-reference table. PDFs written by older software gain the most; a file already using these structures is returned unchanged rather than rebuilt. It does not re-encode or discard images. For significant file size reduction, image re-encoding is required — which cannot be done safely in the browser without potentially altering image quality.
If a PDF is already efficiently structured or its size is dominated by images, structural compression will have little effect. Re-encoding images — which this tool intentionally does not do — is the main way to shrink image-heavy PDFs.
No. The rebuild only changes how the document's objects are stored and indexed, not what they contain. Page content, fonts and images are preserved exactly, and the original Producer and modification date are left in place.
Text-heavy documents written with a plain cross-reference table, which is what older generators and many export pipelines still produce. Those roughly halve. A PDF whose bytes are mostly images barely moves, because the images are left alone.
We'd rather be upfront: structural compression alone won't shrink image-heavy PDFs much. If you need real size reduction for scanned or photo-heavy PDFs, you'd need image re-encoding, which isn't what this tool does.
Not currently on this site. For image-heavy PDFs, consider extracting images, compressing them with Image Optimizer, and rebuilding the PDF.
No. Text remains fully selectable and searchable since the underlying text objects and fonts are untouched.
Yes, but if the file was already optimized (by this tool or another), running it again typically yields little to no further reduction.
