Compress PDF to Target File Size

Lossless mode first tries QPDF object and stream optimization in WebAssembly. If that cannot initialize, the tool attempts a pdf-lib rebuild of an unencrypted document. Already optimized files may gain little or no reduction; there is no fixed percentage saving. The tool keeps the original bytes if the structural result is larger. Keep the source separately, especially when signatures, attachments or form behavior matter, and inspect the exported copy in the receiving workflow.

Automatic and target modes may render each page into a JPEG-backed replacement PDF if structural processing cannot meet the requested limit. This is a bounded search over image quality and resolution. It may fail to reach a tiny target. The tool reports that failure instead of claiming that an oversized result succeeded. Rendering removes selectable text, editable fields and links. It cannot preserve a cryptographic signature merely by keeping a visible signature image. Check readability before submission and use lossless mode if these structures must remain.

A size preset is a convenience, not an agency requirement. The upload field for your specific form controls the accepted size, format and grouping. Department of State NVC scanning guidance specifies a 4MB maximum per document and asks for readable color scans. USCIS workflows can have different requirements; do not transfer a number from another form or another agency. If a document already fits and looks clear, leaving it unchanged may be preferable. Check every downloaded page and keep a margin below a stated byte limit.

Choose a target from the receiving portal's actual requirement. A 200KB preset does not mean USCIS or another agency universally requires 200KB. Use the portal's instructions, permitted formats and document grouping rules. A readable file already below the permitted limit may need no compression at all. The tool interprets KB as 1,024 bytes. Portals may use different units or rounded displays, so keep a margin. If an adaptive attempt remains above the target, the tool reports that it could not meet the limit rather than offering a falsely successful oversized result. Lossless mode can legitimately return a file above the selected target because it prioritizes retaining page content. File size and document acceptability are separate checks you must perform on the downloaded output.

Adaptive compression renders each page and places a JPEG image in a new PDF. That can remove selectable text, links, editable fields, bookmarks and structural accessibility information. A visible signature image is not a preserved cryptographic signature. Keep the original when any of these features matter and use a suitable alternative if the receiving workflow requires them. Review small numbers, handwritten details and color-dependent marks at a normal reading size. Compare the page count and order. Do not assume a barcode is valid merely because its outline looks sharp. A long color scan may be impossible to fit into a very small target while preserving evidence quality. Re-scan efficiently, use a larger permitted target or split only when the receiving instructions allow it. Repeatedly compressing an already degraded copy rarely restores lost detail.

Suppose your original is 700KB and the actual upload limit is 1MB. Keep the readable original unless you have another reason to reduce it. If the limit really is 500KB, first try structural optimization. If that is insufficient, permit adaptive processing and inspect the resulting images carefully. Stop when important text becomes difficult to read, even if a smaller size is technically possible. Save the output under a new name so the original remains available. The compression path runs inside the browser and does not upload your document. WebAssembly and PDF.js assets may need to download first, so ordinary asset requests do not indicate a document upload. On a shared computer, saved outputs require their own storage and deletion decisions after the browser task has finished.