pdfviz

See what makes your PDF large

or drop one here
Choose a PDF first.

See exactly where a PDF uses its bytes without uploading or changing it. The tool follows the document graph and matches images, font programs and content streams to the pages that use them. The report separates images, fonts, page content, metadata, embedded files, old objects and file structure, checks that the totals equal the source file, and names the one change worth making. The PDF stays in this browser tab, with your file never uploaded and no account.

How to find what makes a PDF large

  1. Choose a PDF

    Press Choose a PDF, or drop one on the page. The file stays in this browser tab.

  2. Walk the PDF objects

    The tool follows the document graph and matches images, font programs and content streams to the pages that use them.

  3. Reconcile every byte

    Parsed object sizes are apportioned to the physical file. Headers, indexes and other wrapping stay visible as their own category.

  4. Inspect and act

    Review the largest objects and follow one recommendation based on the part that actually takes the most space.

What the numbers mean

Every number is the stored byte count, not the decoded one. A JPEG can decode into many megabytes of pixels and still store far fewer, so the report always states what the file holds. Object headers, cross-reference data and trailers appear as their own category rather than being spread silently across the others, and the categories plus that wrapper must equal the exact source byte count. When they do not, the run fails instead of reporting.

PDF parsers normalize object syntax and object streams, so parsed sizes are apportioned back to the physical file. Analyze the file size before you compress: the breakdown names the category that holds the bytes, and the table below says whether compression can move it.

What each size category counts

What each size category counts, and whether compression can move it
CategoryWhat it countsCompression moves it
ImagesWhat it countsPhotos, scans and logos at their stored size, not their decoded pixel count. Compression moves it Yes Compress PDF re-encodes placed images at 150 DPI.
FontsWhat it countsEmbedded font programs. A full font costs the same whether a page draws 12 glyphs or 1,200. Compression moves it TrueType only The font arm subsets embedded TrueType to the glyphs the pages draw. Type 1 and CFF programs are left alone.
Page contentWhat it countsThe drawing instructions for each page: text runs, vector art, positions. Compression moves it Rarely A large number here points at vector art rather than text, since text runs are small.
Embedded filesWhat it countsAttachments carried inside the PDF, such as the spreadsheet a chart came from. Compression moves it No Remove the attachment or the bytes stay.
MetadataWhat it countsXMP packets, document info and the producer trail. Compression moves it No Remove PDF metadata clears it for privacy; for size, look elsewhere.
Unreferenced objectsWhat it countsObjects left by earlier saves that the current document no longer reaches. Compression moves it Usually A clean save can drop them. It can also invalidate a digital signature.
PDF structureWhat it countsThe page tree, outlines, annotations and form field definitions. Compression moves it No A fixed cost of what the document is, not of how it was saved.
File wrapperWhat it countsObject headers, cross-reference data and trailers. Compression moves it No Leftover wrapping the parser cannot apportion: headers, cross-reference data, trailers.

The breakdown reports the split. It never rewrites the PDF.

When to break down a PDF before compressing

A size breakdown is the diagnostic step before a fix. When a PDF is too large to email or too slow to open on a phone, the question is what is taking the space. A report with one 40 MB image and 2 MB of text needs a different fix from one with 1 MB of images and 40 MB of embedded fonts. The breakdown names the category, and the recommendation points at the single change that moves the most bytes.

For oversized images, Compress PDF re-encodes placed images at a target DPI. For a font embedded in full where a subset would do, rebuilding the document from its source is the fix. Run the breakdown again after the fix to confirm the total dropped.

What to do with the answer

A breakdown is a diagnosis, not a fix. Once you know which images, fonts or content streams carry the weight, How to compress a PDF covers the ordinary route, and Compress a PDF without losing quality covers what can be reduced with no visible change and what cannot.

Questions about PDF size breakdown

Does the size breakdown change my PDF?
No. This tool reads the file and produces a report. It does not write a new PDF.
Do the categories add up to the exact file size?
Yes. The receipt checks that the categories and file wrapper equal the source byte count exactly. When they do not, the run fails instead of reporting a wrong total.
Why is file wrapper a category?
A PDF also contains object headers, cross-reference data, trailers and other physical wrapping. Parsed objects do not retain every original byte offset, so that difference is shown instead of hidden.
Can it tell me which image is large?
Yes. The largest-object list includes page location, pixel dimensions, placed DPI and codec when that information is available.
Does the size breakdown upload my PDF?
No. The PDF object graph, page matches and size report stay in this browser tab.
What should I do with the result?
Start with the single change the report names. That is usually one oversized image, or a font embedded in full where a subset would do. Compress PDF acts on the first case; rebuilding the document from its source acts on the second. Run the breakdown again afterwards to watch the categories move.
Why does my PDF show a large page content category?
Page content streams are the drawing instructions for each page. A large value usually means vector art, not text, since text runs are small. These streams are already compressed, so compression rarely moves them.
Can I see the breakdown for a password-protected PDF?
Unlock the PDF first. The tool reads the object graph, which requires access to the unencrypted structure.

Keep working with your PDF

This breakdown reader runs in your browser: your PDF is not uploaded, and the report changes nothing in the file. The tools below each run in their own browser tab.

Where to go next