One file at a time
✓ Two-stage engine: tries lossless PDF structure optimisation first. If that’s not enough, only image-heavy pages are rasterized — text pages are always copied losslessly (except Maximum).
Stage 1 — Structure optimisation: the engine strips embedded XMP metadata, page thumbnails, and editor private data, then re-saves the PDF with compressed object streams. This alone can reduce bloated or scanner-exported PDFs by 15–30% without touching any content — text stays fully searchable.
Stage 2 — Smart per-page compression: each page is classified by how much of its area is covered by raster images. Text-only pages are never recompressed. Pages that are mostly image (scans, photos) are re-rendered at the preset resolution. Mixed pages — text with logos or figures — keep their text layer exactly; only the embedded images are recompressed in place, never flattened.
Low Compression — images at 200 DPI, JPEG 85%. Print quality. Text pages are always lossless.
Recommended — images at 150 DPI, JPEG 72%. Smart balance: text lossless, image pages compressed well. Good for email and Drive.
Maximum Compression — all pages rasterized at 72 DPI, JPEG 50%, regardless of content. Smallest file. Text becomes non-searchable. Good for WhatsApp (16 MB limit), IRCTC, DigiLocker.
The engine always returns whichever result is smaller (Stage 1 or Stage 2). If compression would make the file larger, the original is returned unchanged.
Will my text still be searchable?
Yes — on Low and Recommended presets, text-heavy pages are always copied losslessly. Only image pages are rasterized. Maximum Compression rasterizes all pages, making text non-searchable.
Why did my PDF barely compress?
Text and vector PDFs have little to compress — the stage 1 structure pass will still trim some overhead. If you need a smaller file, try Maximum Compression (text becomes non-searchable). Alternatively, split the PDF and compress parts separately.
Why is my scanned PDF now tiny?
Scanned PDFs contain large embedded images (often 300+ DPI). Re-rendering at 1.2× or 1.5× scale compresses these dramatically. The output is still sharp at normal viewing sizes.
The tool says "already optimised" — what does that mean?
Both Stage 1 and Stage 2 produced a file at least as large as the original. This happens with PDFs that were already fully optimised (e.g. output from ilovepdf or Acrobat's Optimizer). The original is returned unchanged.
Is there a file size limit?
No server-side limit. Large files (100 MB+) may be slow — each page is processed individually so even 500-page PDFs complete without crashing.
PDF still too large after Maximum Compression?
Split the PDF using the Split PDF tool, then compress each part separately.
Need a PDF under exactly 1MB?
See our guide: Compress PDF below 1MB — for IRCTC, DigiLocker and visa uploads →
Need to compress for WhatsApp?
See our guide: Compress PDF for WhatsApp — tips for reliable delivery on any device →
Need to compress for email?
See our guide: Compress PDF for Email — fit Gmail and Outlook attachment limits →
Your files never leave your browser. All compression work happens inside your device's memory — no file is sent to any server, no copy is retained, and nothing is transmitted over the network. Sensitive documents such as tax filings, bank statements, and ID proofs are processed locally and discarded the moment you close or refresh the tab.
No account is required. There is no sign-up form to fill out, no email to verify, and no subscription to manage. Open the page, drop your file, and download the result — the full compression engine is available to every visitor on every visit, with no watermarks and no export limits attached to a free tier.
LovelyPDF works on any device with a modern web browser. Whether you are using Chrome on Android, Safari on an iPhone, or Edge on a Windows laptop, the entire compression process runs client-side with no app to install and no operating system restrictions.
Results depend on content type. Scanned documents with high-resolution images can often shrink by 60–85% on the Recommended preset. PDFs composed mostly of text and vector graphics typically compress by 5–15% because there is little image data to reduce. Mixed-content PDFs — such as reports with charts and photographs — usually land in the 25–50% range. If you need to hit a specific target like 1 MB, switch to Maximum Compression: it rasterises every page at low quality for the smallest possible output.
On the Low and Recommended presets, text-only pages are copied losslessly — the text layer, fonts, and vector graphics remain identical to the original. Only pages that contain embedded images are re-rendered at the preset JPEG quality. The Recommended preset (JPEG 65%) produces output that looks sharp at normal screen and print sizes. Maximum Compression rasterises every page including text, converting real characters into a pixel image — the file looks the same on screen but text is no longer selectable or searchable.
Yes, there is no file size limit. The tool processes each page individually inside your browser, so even a 200-page, 100 MB scanned document will complete without crashing. Processing time scales with page count and image resolution — a 50-page scanned file typically takes 30–90 seconds depending on your device speed. If the browser appears unresponsive while the progress bar is moving, that is normal behaviour; the tab will respond once processing is complete.
This message appears when both the lossless structure pass and the rasterisation pass produce a result no smaller than the original. It typically means the PDF was already fully optimised — for example, a file previously processed by Adobe Acrobat's PDF Optimizer, iLovePDF, or another compression tool. The original is returned unchanged in this case. Switching to Maximum Compression can sometimes still reduce size further by rasterising at very low quality, though the trade-off is that text will no longer be searchable.