Here, 1 MB = 1,024 KB = 1,048,576 bytes. Choose a target from 1 MB to below 25 MB; the exact byte ceiling appears under the controls. Switching MB and KB keeps the same target. Results can be smaller than 1 MB: this is the minimum target you can select, not a minimum output size. Keep your receiving form’s exact byte limit in mind.
Existing extracted text is compared with the result. A scanned PDF without a text layer remains a scan. Read the text warning and open every output page before submitting; a size fit alone does not confirm readability.
The tool reports failure and shows the smallest tested result. Keep the original, reduce unnecessary pages if the receiving service allows it, or ask that service for a larger allowance. Editable fields and links can be flattened by compression.
Your PDF is read and processed in this browser. The tool has no PDF upload endpoint, account, analytics or advertising code. The app files are downloaded from the hosting provider, which may handle normal connection logs. A private preview also uses its provider’s access controls.
The offline cache stores the app’s static files. It does not store your chosen PDFs or generated results. Files remain in memory while processing. Downloaded PDFs are saved wherever your device’s browser places downloads. Close the tab when finished on a shared device.
Compression turns pages into images and reconstructs supported extracted text. It does not perform OCR. Links, editable fields and vector graphics can be flattened. Non-Latin characters, text positions, rotation and reading order can change. Matching extracted text does not establish visual or semantic fidelity, accessibility, valid signatures or suitability for an official submission.
Files are limited to 25 MB (26,214,400 bytes) and 50 pages. Password-protected and malformed PDFs are rejected. Pages that would render above 4 million pixels or 4,096 pixels on an edge are rejected before canvas allocation. Complex files may still exceed a phone’s available memory even within these limits. Review every page before submitting a result, and keep the original.
Six fixed image settings are tried from highest to lowest detail. The first output below the byte limit is offered; this does not guarantee the best possible compression or readable output at every target. A failed target is reported explicitly.
Send feedback opens an external Google Form. Only information you submit in that form is sent to Google and made available to the form owner. Do not submit PDFs or sensitive document details. Device/browser and reply email are optional. The form does not automatically collect email addresses; Google may show you your own signed-in account and apply its own privacy terms. Feedback notifications are enabled for the owner. PDF processing does not send files to this form.
Use synthetic documents during this private validation. Public release, optional analytics and advertising each require review before introduction.