Identify what makes the PDF large
Start by opening the document at 100% zoom and checking several representative pages. Try selecting a sentence. If you can select individual words, the PDF probably contains text objects; if the entire page behaves like one picture, it is likely a scan. Mixed documents need both kinds of checks. Photograph-heavy reports usually offer more compression opportunities than simple invoices.
Keep the original untouched and note its file size. Decide what the smaller copy must support: email delivery, screen reading, or printing. A strict upload limit is a constraint, not a quality target. Do not keep reducing quality after the document already fits comfortably.
Set up a local compression workflow
On Windows, install PDF24 Creator from its official website. Its desktop tools process files locally, unlike a browser upload workflow. This distinction matters for contracts, medical records, and internal reports. Use a working folder with the original, trial outputs, and a clearly named final copy.
Open the compression tool, add one PDF, and choose a separate destination. Settings and labels can vary between releases, so look for controls covering image resolution, image quality, and color. Begin with a conservative preset or modest reductions. Compress one representative file before processing an entire folder.
Choose resolution and color deliberately
For screen-oriented documents, try approximately 150 dpi as an initial image-resolution target, then inspect the result. Small scanned type, engineering drawings, and tiny labels may need 200–300 dpi or the original resolution. These are starting points rather than guarantees. Increasing resolution cannot restore detail missing from the source scan.
Color conversion can save space when color is not meaningful. Grayscale often works for receipts and ordinary correspondence, but colored annotations, heatmaps, and signatures may carry information. Black-and-white thresholding is riskier: faint pencil marks and thin punctuation can disappear. Avoid applying it indiscriminately to mixed pages.
| Content | Initial approach | Main risk |
|---|---|---|
| Digital text | Preserve text objects | Rasterizing sharp text |
| Scanned letters | Moderate downsampling | Blurred small characters |
| Photos | Adjust image quality | Blocking and banding |
| Line drawings | Retain detail | Vanishing fine lines |
Run a controlled comparison
Export a first trial, then compare it with the original side by side at the same zoom. Check the smallest text, a dense table, a photo, and a page containing thin lines. Enlarge suspicious areas to distinguish normal screen rendering from actual damage. If printing matters, print one demanding page rather than judging only on a monitor.
Record the original size, compressed size, and settings in a short note. Change one variable for the next trial: image quality, resolution, or color mode. This makes it easier to identify the source of damage. Keep the smallest version that passes your checks, not simply the smallest version produced.
Preserve searchability and document behavior
After compression, search for a known phrase and copy a paragraph into a text editor. Compression is not OCR, and it will not make an image-only scan searchable by itself. If searchable scans are required, add OCR in a separate, tested step and verify the resulting text layer before delivery.
Test bookmarks, clickable links, page orientation, and form fields where applicable. Rewriting a signed PDF can invalidate its digital signature even when the pages look unchanged. Preserve the signed original and obtain an approved distribution copy instead of treating compression as a harmless edit. For accessible documents, visual inspection alone does not establish that tags or reading order survived.
Avoid common size-reduction mistakes
Do not print to PDF as your default compression method. That route can flatten forms, remove document structure, or change links. Do not convert every page into a JPEG just to simplify processing: digital text becomes an image, which can reduce both readability and accessibility. Repeatedly recompressing an already damaged output also compounds image degradation.
Removing pages is useful only when those pages are genuinely unnecessary and removal is authorized. Cropping a page usually changes what is visible rather than securely deleting hidden content or guaranteeing substantial size savings. Redaction and metadata removal are separate tasks. If sensitive information is involved, use an appropriate redaction process and inspect the exported result.
Use a final delivery checklist and official references
Before sending, confirm that the final file opens in a second PDF viewer, contains the expected page count, and stays below the actual upload limit. Inspect the beginning, middle, and end rather than trusting a successful export message. Give the final copy a distinct filename so recipients and colleagues do not confuse it with the master.
Checklist: readable small text; intact diagrams; accurate page order; working search where expected; acceptable print sample; preserved original; verified destination. If no reasonable setting meets the size limit, ask whether multiple files or a secure download link is acceptable instead of silently destroying detail.
Use the official PDF24 compression guidance to understand the task, and download the offline application from the Creator page. Explore other no-cost utilities in freebies, checking each tool’s privacy model before uploading documents.