A scan stores the page as a picture, often much larger than the words it shows. Compress PDF offers two ways to make a smaller copy. The right result is the smallest file that still meets the recipient’s reading needs.
1. Check the file and the upload limit
Keep the original. Note its size and the receiving site’s limit, then try selecting a word in a PDF reader. If the whole page behaves like a picture, it is probably an image-only scan. A scan with OCR can also have selectable text.
The tool accepts a PDF up to 32 MB. It has no exact target-size setting: choosing Screen copy does not guarantee a file below 1 MB or any other threshold.
2. Choose a mode for the intended use
- Lossless: cleans up PDF structure without turning page content into pictures. Start here when searchable text, crisp vector content or interactive fields matter.
- Screen copy: also tries rendering all pages as lower-resolution JPEG images. This can help with image-heavy scans. If that version is the smallest, searchable text and form fields become part of the image.
Both modes keep the smallest candidate, including the original. A PDF that is already compact may be returned unchanged. Screen copy can also return a lossless version if that is smaller than the image version.
3. Compare two outputs from the same original
- Choose your original and run Lossless. Download the result and note its size.
- Choose the original again and run Screen copy. Download that result under a distinct name.
- Open both results and inspect the same difficult area: a small receipt total, faint date, stamp or dense table.
For example, a receipt may look fine as a page thumbnail while its decimal point becomes hard to read when enlarged. Compare that amount with the source. If a barcode or QR code matters, test it with the reader the recipient will use. Check text selection and form fields if you need those features.
4. Inspect two measured examples
These files were created for a synthetic demonstration and processed by the hosted tool on November 4, 2026. Their savings are not typical or guaranteed.
- Screen copy: an image-heavy three-page report went from 6,721,599 bytes (original PDF) to 57,853 bytes (result PDF). The fine print is visibly softer and its selectable text is gone. This illustrates the image-quality tradeoff; it is not a real customer scan.
- Lossless: an eight-page checklist went from 45,675 bytes (original PDF) to 8,219 bytes (result PDF). Its text was preserved and page renderings matched at 72 DPI. The original was deliberately generated without content-stream compression, leaving unusual room for savings.
5. If the PDF is still too large
Repeatedly compressing the output does not guarantee further savings. Return to the original and consider removing unnecessary pages. If separate files are accepted, split the PDF into smaller documents. For a new scan, use the scanner’s resolution and color controls, then check the smallest print before scanning the whole batch.
Keep a higher-quality original when the document will be printed or archived. If you need to restore search after flattening, OCR PDF can add recognized text, but recognition needs checking and may increase file size.
Hosted compression runs on our server; uploaded files are removed after each operation. Download the copy before leaving the tool. Read file privacy or compress a scanned PDF.