What makes a PDF large
A PDF can contain page instructions, embedded fonts, photographs, scanned pages and document information. A long text document may be quite small, while one page containing a high-quality photo can occupy several megabytes. The page count alone therefore does not tell you how much space an optimizer can save. This tool rewrites the PDF using compressed object streams and optionally recompresses compatible embedded JPEG images. It does not flatten every page into a new image.
Select files saved on your device and choose whether image recompression is appropriate. Files are processed locally in a worker rather than uploaded to a server. A batch is limited to ten files, two hundred MiB per file and five hundred MiB in total. Each document may contain up to one thousand pages. Large documents need substantial memory even when their compressed file size is within these limits; a desktop browser is recommended, and processing can still fail on memory-constrained devices. A MiB is 1,048,576 bytes; external websites may use a different meaning for MB. JPEG source images up to fifty million pixels can be processed when the selected pixel limit brings the decoded output below sixteen million pixels. Without downsizing, images above sixteen million pixels are left untouched to limit memory usage.
Choose quality for the actual document
The default JPEG quality is 75, and the default image long-edge limit is 2,339 pixels. This is approximately 200 DPI for an A4 full-page scan. The 3,508-pixel setting is approximately 300 DPI on A4, while 1,754 pixels gives approximately 150 DPI. These are pixel caps, not measured placement DPI: a smaller picture on a page or a different paper size has a different effective resolution. Images keep their aspect ratio, small images are not enlarged, and PDF page dimensions remain unchanged. Choose original resolution if reducing image pixels is not appropriate. Lowering it can help with photographs, but it can also soften lettering inside scanned pages. Compare small type, barcodes, thin rules and coloured diagrams in a PDF reader. Keep the original document so you can return to it if the result is unsuitable. Turning off image recompression leaves image bytes unchanged and only attempts structural optimization; an already compact PDF may show no improvement.
Only supported RGB and grayscale JPEG streams are re-encoded. The output can store grayscale pictures as RGB JPEGs, with the PDF colour-space entry updated to match. Images with colour profiles represented by PDF colour spaces, masks, explicit decoding transforms or other unsupported settings are skipped conservatively. Other image encodings and embedded fonts are not recompressed. This makes the operation less aggressive than a full page-rasterization compressor, but preserves existing text and vector content without converting them to pixels.
Check before submitting
Each completed file shows its original and resulting size and has a separate download link. If rewriting does not reduce the size, the original bytes are offered instead. This is not a guarantee that the result meets a particular upload limit. There is no target-size mode or page deletion. Image resolution is only reduced according to the selected pixel limit. Verify the receiving system’s byte allowance and inspect all important pages before submitting.
Encrypted PDFs and documents containing signature fields are rejected. Do not use an optimized copy as a substitute for an authenticated original. Although the optimizer does not intentionally remove forms, links or annotations, complex interactive documents should be checked in their intended reader. Archival compliance, accessibility conformance and preservation of specialised document features are not certified by this tool.