How this compressor works
Drop in a PDF and each page is redrawn right in your browser using pdf.js, the same rendering library Firefox uses to display PDFs on screen. Each rendered page is then re-encoded as an optimized JPEG, and the document is rebuilt from those images. The whole pipeline runs in a Web Worker on your own machine — nothing goes up to a server and nothing extra comes down. SafeFileConvert keeps the rest out of the way too: no account, no daily task counter, and no watermark on the result.
Best for scans, with one trade-off for text PDFs
Because every page comes out as an image, the output is no longer selectable or searchable text. For scanned paperwork — the files this tool is mostly used on — that changes nothing, since a scan is already a picture of a page. For a PDF exported from Word or Google Docs, keep the original for editing and reading, and treat the compressed copy as an upload version only. One more safeguard: if re-encoding would actually grow a small text-only file, you get your original back untouched instead of a bigger 'compressed' one.
Hitting a portal's size limit
Most people arrive here because a form said no: job-application systems, email attachments, and government uploads all enforce ceilings, and many state them in kilobytes. Set a target and the compressor makes several passes at stepped-down resolution and quality, stopping as soon as the file lands under your number — under it, not at it, since a file that grazes the limit can still bounce. If a target is unrealistic at readable quality, you receive the closest smaller result and a plain notice saying the goal was out of reach. Dedicated preset pages exist for 100 KB, 200 KB, and 500 KB. Encrypted PDFs need their password removed first; the tool tells you so instead of failing quietly.