How to Compress a PDF in Your Browser Without Uploading It

A PDF that is too large for an email or upload form creates a simple-sounding problem: make the number smaller. The difficult part is deciding what the file is allowed to lose. A typed report, a fillable application and a phone-scanned packet may look similar in a viewer, but they contain very different things. Treating all three as a stack of pictures can reduce bytes while quietly discarding the features that made the document useful.

The Compress PDF tool therefore offers two deliberate paths. Lossless cleanup rewrites the PDF structure without turning pages into images. Raster pages renders each page as JPEG and rebuilds a visual copy. Both happen in your browser, but they solve different problems and produce different kinds of output.

The ISO 32000-2 PDF standard defines PDF as a format for exchanging and viewing electronic documents independently of the environment that created them. In practice, one file can combine text, vector paths, photographs, links, form controls, annotations, metadata and structural information. File size is the sum of those parts, not a direct measure of page count or visual complexity.

That is why two forty-page documents can behave very differently. A report made mostly from text and vector lines may already be compact. A six-page scan can be much larger because every page is a high-resolution photograph. There is no universal compression percentage that can be promised before the file is inspected, and a safe rewrite can occasionally be the same size or slightly larger.

Lossless mode loads the document and saves a fresh working copy with PDF object streams enabled. It does not resample page images or flatten the page into pixels. The page content, selectable text, links and form structure remain available, and the page count does not change. The source on disk is never overwritten.

The benefit depends entirely on the input. Files assembled by older software, repeatedly edited copies and documents with inefficient internal structure may shrink. A modern exported PDF whose objects are already compressed may not. When the output is not smaller, the tool says so instead of presenting a larger file as a successful optimization.

Raster mode renders every page at the selected resolution — 72, 96, 120 or 150 DPI — encodes that rendering as JPEG at the selected quality, and places the result on a new PDF page with the same physical dimensions. A scan that contains large photographs often becomes much smaller because its pages are encoded once at a controlled resolution and quality.

The price is explicit: each output page is one image. Original text is no longer selectable or searchable, links stop being links, fields can no longer be filled, annotations are baked into the rendered appearance, and document tags are gone. Running OCR afterwards can add a new invisible searchable text layer, but it cannot restore the original links, fields or tagged structure.

1. 1

   Keep the original

   Work from a copy when the document is important. Compression creates a new download, but a clearly named original makes comparison and recovery straightforward.

2. 2

   Try lossless cleanup

   Open the PDF, keep Lossless cleanup selected and run the tool. Compare the input and output byte counts in the export report.

3. 3

   Open the result

   Check several pages, select a sentence, follow an important link and test a form field if the document has one. A smaller number is useful only when the document still does its job.

4. 4

   Escalate only if needed

   If the file is still above the limit and the document can safely become image-only, switch to Raster pages. Start at 120 DPI with a moderate JPEG quality rather than choosing the lowest values immediately.

5. 5

   Compare at normal zoom

   Inspect small type, signatures, fine table rules and pale scans at 100% zoom. If they soften too much, raise DPI or quality and run another local copy.

6. 6

   Verify the receiver's constraint

   Confirm the final byte count is below the email, portal or document-management limit, then retain the smallest version that remains readable and fit for purpose.

DPI controls how many pixels are used to represent the physical page. Lower DPI removes spatial detail, so thin strokes and small characters have fewer pixels available. JPEG quality controls how aggressively those pixels are encoded. Lower quality can introduce blockiness or halos even when the page dimensions stay the same. They are separate levers, and lowering both at once makes it hard to tell which one damaged the result.

• 72 or 96 DPI: a last-resort viewing copy for screens, especially risky for small type and fine tables.
• 120 DPI: a useful first raster test for ordinary office scans when attachment size matters.
• 150 DPI: more detail for signatures, stamps and smaller print, usually with a larger output.
• JPEG quality: adjust in small steps after choosing DPI; a photo-heavy scan tolerates lower quality better than a page of crisp type.

The selected file is read into the tab and the compression request runs in a local worker so page processing does not require an upload endpoint. The result and its before-and-after report are returned to the page for download. There is no account, server queue or cloud copy created by the tool. Closing the tab releases the working data unless you explicitly save the output on your device.

Large image-heavy documents still use real memory and processor time on your device. If a high-DPI raster attempt strains a phone or older laptop, lower the DPI, process on a desktop, or split the document into smaller sections and merge the finished pieces afterwards.