Select your PDF
Choose one PDF from your device. The browser opens it locally and shows the page count and a first-page preview.
Reduce PDF file size directly in your browser. Choose the balance between document quality and file size that works best for your document.
Choose a PDF from your device or drag it here.
Compression happens locally in your browser. Your document is not uploaded to our server.
Preparing document…
Choose one PDF from your device. The browser opens it locally and shows the page count and a first-page preview.
Start with Recommended, or choose Strong or Maximum when a smaller file is more important than fine visual detail.
Click Compress PDF. The browser rebuilds the pages, validates the result, and compares the actual output bytes with the original before offering a download.
There is no single compression setting that is ideal for every PDF. A scanned form, a presentation full of photographs, and a text-only research paper contain very different data. This tool therefore offers three practical image-based levels rather than pretending that one button can reduce every file by the same percentage.
Recommended targets roughly 150 DPI with relatively high JPEG quality. It is intended as the first choice for ordinary documents where readable text and graphics still matter. If the source is already compact, this level may produce little or no useful reduction.
Strong targets roughly 120 DPI and uses more aggressive JPEG compression. It can work well for scanned documents, reports with photographs, and files that need to fit an upload or sharing limit. Fine image detail can become softer.
Maximum targets roughly 96 DPI with substantially lower JPEG quality. It prioritizes file size. Use it when the document only needs to remain visually readable and you accept a larger loss of fine detail.
A PDF can contain selectable text, fonts, vector drawings, photographs, scanned pages, annotations, forms and already-compressed data streams. That is why a truthful PDF compressor cannot guarantee a fixed reduction. In this implementation, PDF.js renders each visible page at a controlled resolution. The page is placed on a white canvas, encoded as a JPEG, and embedded into a new PDF page with the same apparent page dimensions and orientation.
This is genuine lossy compression. It often helps image-heavy or scanned PDFs because high-resolution page imagery can be recreated at a lower resolution and JPEG quality. However, a small digital-born PDF may already store text and vectors extremely efficiently. Rasterizing such a file can create a larger output. When that happens, this tool reports the result honestly and recommends keeping the original.
Lossless optimization changes PDF structure without intentionally changing the visible page. It may compress streams, remove unused objects, or reorganize internal data. Savings can be modest when images already dominate the file. Lossy compression changes image representation or resolution to save more space. The three production modes on this page use lossy image-based recompression; they are not described as lossless.
The tool does not expose a “Preserve Text & Vectors” mode because no browser-side structural optimizer is bundled in this release. Avoiding a fake lossless option is better than showing a mode that merely loads and re-saves the same PDF.
It can. Recommended keeps more image detail than Strong, and Strong keeps more than Maximum. The exact visible difference depends on the source. A large photograph can tolerate more compression than a tiny diagram with fine labels. Always open the downloaded PDF and check important pages before deleting or replacing the original.
Many text-heavy PDFs are already compact because text instructions and vector paths need very little storage. Photographs may already be saved as compressed JPEG streams, and exported PDFs may already contain optimized object streams. In these cases, another raster pass can add overhead or produce a larger file. This tool measures the candidate instead of inventing a saving. If the candidate is larger, you receive a neutral “No smaller result” message and a button to keep the original.
Scanned documents are often good candidates for image recompression because each page may contain a large high-resolution image. Start with Recommended and inspect small text, signatures and stamps. If the file is still too large, try Strong. Maximum is best reserved for cases where meeting a file-size limit matters more than preserving fine detail. The optional grayscale setting can help some black-and-white paperwork, but it permanently removes color.
If you need a smaller attachment or a document that fits a portal limit, first check the required maximum file size. Use Recommended before moving to Strong or Maximum. Do not rely only on the displayed reduction percentage: open the resulting PDF and confirm names, numbers, diagrams and signatures are still readable. For forms that must remain fillable or links that must stay clickable, image-based compression may not be the right workflow.
This Compress PDF implementation reads and processes your document in the browser. The PDF bytes are not posted to an I Love PDF Tools upload endpoint. The page does load the pinned PDF.js processing library as a static browser dependency, but your selected PDF content and filename are not sent to that library host by the application.
Because processing is local, performance depends on your device, browser, PDF complexity and available memory. Very large pages are automatically rendered at a lower effective scale when necessary to reduce the chance of exhausting browser memory, especially on mobile devices.
Make image-heavy documents easier to attach and transfer.
Reduce scans, reports or appendices before uploading to submission portals.
Shrink scanned certificates or portfolio PDFs while checking that text stays readable.
Reduce report, invoice and presentation file sizes for practical sharing and storage.
The source may already be optimized or mostly text/vector content. The tool will recommend the original rather than calling the larger candidate a success.
Try Recommended or raise the resolution/quality in Advanced settings. Maximum deliberately uses lower resolution and JPEG quality.
Try Strong or Maximum, close other memory-heavy tabs, or use a desktop browser. The tool already caps page canvas size, but extremely large or complex documents can still exceed device limits.
Unlock the document first with the Unlock PDF tool, then try compression again. This tool does not crack passwords.
That is an expected limitation of raster compression. Pages are recreated as images, so interactive elements can be flattened.
This tool is designed to compress a PDF in the browser without requiring an account or paid conversion API.
No PDF upload endpoint is used by this implementation. Your selected document is read and processed locally in the browser.
Text-heavy, small or already optimized PDFs can be extremely efficient. Recreating them as page images may not reduce their size.
The three modes are lossy. Recommended keeps more visual detail than Strong, and Maximum trades more quality for a smaller practical result.
They use different target resolution and JPEG-quality settings. Strong and Maximum render fewer pixels and use more aggressive JPEG compression.
Yes. Scanned and image-heavy PDFs are often the documents most likely to benefit from this raster recompression approach.
Not reliably. Raster compression recreates each page as an image, so selectable/searchable text is normally flattened unless a separate OCR text layer is added.
Yes in a current capable mobile browser. Mobile devices have tighter memory limits, so the tool automatically uses conservative canvas limits.
This release does not process protected PDFs. Unlock the file first and then compress the unlocked copy.
The original may already be highly optimized. The tool compares actual byte sizes and recommends keeping the original when the candidate is larger.
There is no honest universal size limit. Practical capacity depends on page count, page dimensions, image complexity, browser and available device memory.
The browser-held file, generated Blob URLs and processing state disappear with the page session. The application does not upload your PDF to its server.