Most PDF size reduction online is image compression: pages are re-rendered as pictures at a lower quality. That works — it is exactly what our Compress PDF tool does, and for scanned documents it cuts 60–80% — but it has a real cost. Text stops being selectable and searchable, hyperlinks die, and form fields flatten. For a contract someone needs to search, a form someone needs to fill, or a report someone needs to quote from, that trade-off is unacceptable. Optimization exists for those files.
LuraPDF's optimizer rebuilds your PDF losslessly using pdf-lib: the document's internal object structure is rewritten with compressed object streams — a standard PDF feature many older generators never used — and document metadata can be stripped in the same pass. The content layer is carried over byte-for-byte. The savings are honest and modest: typically 5–20% on files with inefficient structure, sometimes nothing on files that were already well packed — and the tool tells you when that is the case instead of pretending otherwise.
Optimization solves a different problem than compression: making a file lighter without changing what it is. These are the situations where that distinction matters.
Legal documents must stay searchable, quotable, and unaltered. Lossless optimization trims file size for filing portals and email without changing a single character — and without the integrity questions that re-rendered pages can raise.
Image-based compression flattens form fields permanently. Optimization is the only size reduction that keeps a fillable PDF fillable — apply it before distributing application forms, intake forms, or surveys.
Quarterly reports, research papers, and documentation are used by readers who search, select, and copy text. Optimization keeps all of that working while making the attachment lighter.
The metadata toggle clears author names, creation dates, and producer software from the file — useful when sending documents outside your organization without revealing internal details.
Screen readers depend on the PDF's text layer. Compression destroys it; optimization preserves it completely. For documents that must remain accessible, optimization is the safe choice.
Long-term archives need files that remain exactly what they were. Optimization reduces storage cost without altering content, so the archived copy stays a faithful original rather than a lossy derivative.
A lossless optimizer that runs entirely in your browser, paired with an honest account of what it can and cannot do.
A PDF file is a collection of numbered objects — pages, fonts, images, annotations — plus a cross-reference table that records the byte position of every object so readers can jump straight to what they need. Older PDF generators write each object and its bookkeeping uncompressed, which wastes space; files edited repeatedly accumulate further structural slack. LuraPDF's optimizer loads the document with pdf-lib and writes a brand-new file using object streams, a PDF 1.5 feature that groups non-stream objects together and compresses them as a unit. The content of every object — the actual text, image data, and vector drawing commands — is carried over unchanged.
If you enable metadata stripping, the optimizer also clears the document information dictionary: title, author, subject, keywords, creator, producer, and the creation and modification dates. This both saves a little space and removes provenance details you may not want to share. Everything runs in client-side JavaScript inside your browser tab — no upload, no server, no queue. Because nothing is rasterized, optimization is fast and memory-light even on mobile devices, and the result is a conformant PDF that behaves identically to the original in every reader.
| Property | Optimize PDF (lossless) | Compress PDF (smallest size) | Server-based optimizers |
|---|---|---|---|
| Text stays selectable & searchable | Yes — always | No — pages become images | Varies by tool |
| Typical size reduction | 5–20% (structure only) | 60–80% on scans | Varies — often image-based |
| Form fields & links preserved | Yes | No — flattened | Varies by tool |
| File stays on your device | Yes — browser-only | Yes — browser-only | No — upload required |
Optimization works best when you pick it for the right file. Match the tool to the job with these guidelines.
Use optimization for documents where text must stay selectable, searchable, or fillable — contracts, forms, reports, and anything accessibility-critical
Use Compress PDF's Smallest-size mode for scanned documents instead — scans are already images, so rasterizing costs nothing and saves 60–80%
Strip metadata when sending files externally — author names, creation dates, and producer software are visible to anyone who checks the document properties
Don't expect miracles on modern files — PDFs exported by recent versions of Word, Chrome, or InDesign are often already structurally efficient, and the tool will tell you when no further reduction is possible
Old files shrink the most — documents produced by legacy generators or edited across many sessions carry the structural slack that optimization removes
Keep the original if document properties matter — metadata stripping is irreversible in the output file, though your untouched source file always remains on your device
Drop your PDF above and rebuild it losslessly in your browser: a more efficient internal structure, optional metadata removal, and every word, link, and form field exactly where you left it. No upload, no account, no watermark. If your file is a scan and you need it dramatically smaller, Compress PDF is one click away.