Find out why your PDF is so big
A PDF that should be a few hundred kilobytes turns out to be 40 MB, and the email bounces. Guessing what to fix wastes time. The PDF Size Analyzer opens the file, measures every object inside it and shows exactly where the bytes go: images, embedded fonts, page content, attachments, metadata, thumbnails, colour profiles and objects that are no longer used at all.
You get a size breakdown chart, a list of the five biggest objects, practical advice for this particular file, and a sortable table of every stream with its dimensions, compression method and the pages that use it. It is meant for anyone who needs to understand a heavy PDF before shrinking it — office staff, students hitting upload limits, designers checking print files and developers debugging generated documents.
The analysis runs in your browser with the open-source pdf-lib parser. Your PDF is never uploaded, so confidential reports and scanned IDs stay on your device.
How to use the PDF Size Analyzer
- Drop a PDF onto the upload area or click it to choose one.
- Read the headline and the chart: they show which category takes the most space and its share of the file.
- Check “Biggest objects” — often one or two oversized images or a fully embedded font explain most of the size.
- Use the “Show” menu and the column headings to filter and sort the table of streams.
- Follow the suggestions under the chart, or download the full breakdown with “Download CSV”.
PDF Size Analyzer features
Every byte attributed
The categories always add up to the exact file size. Streams are measured precisely; structure such as dictionaries and the cross-reference table makes up the remainder.
Image details
Each image shows its pixel size, colour space, bit depth, compression filter (JPEG, Flate, JPEG 2000, CCITT, JBIG2) and the pages it appears on.
Font programs found
Embedded TrueType, Type 1 and CFF font files are listed with their names, and full embeds are told apart from subsets.
Hidden weight revealed
Spots attachments, page thumbnails, XMP metadata, ICC colour profiles, unreferenced leftover objects and incremental-save revisions.
Sortable, exportable table
Sort by size, page or object number, filter by category, then copy the table or save it as CSV for a spreadsheet.
Private by design
The file is parsed locally in your browser. Nothing is sent to a server.
What usually makes a PDF large
| Cause | Typical sign in the analyzer | What to do |
|---|---|---|
| High-resolution photos or scans | Images take 60–95%; pixel sizes far above what the page needs | Downsample and recompress images, or rasterise scanned pages |
| Images stored losslessly | Large images with FlateDecode rather than DCTDecode (JPEG) | Convert photos to JPEG; keep lossless only for line art |
| Fully embedded fonts | Font programs of hundreds of KB marked “full font” | Re-export with font subsetting switched on |
| Attachments | An “Attachments” slice in the chart | Remove files the reader doesn’t need, or send them separately |
| Edits saved incrementally | More than one saved revision; large structure share | Use a full “Save As” to rewrite the file cleanly |
| Embedded thumbnails | A Thumbnails category | Drop them — viewers generate previews themselves |
How the measurement works
A PDF is a collection of numbered objects. Large data — image pixels, font files, page drawing commands, attached files — lives in streams, and the analyzer reports each stream’s stored size including its dictionary. It then follows references from the document catalogue to decide what each stream is: an object referenced as /FontFile2 is a TrueType program, one listed under a page’s /Thumb key is a thumbnail, and an object that nothing reaches from the catalogue is marked as unused.
Small objects such as page dictionaries are often packed together inside compressed object streams, so their individual sizes cannot be separated. Their combined share, along with the cross-reference table and any earlier revisions of edited objects, is shown as “Structure & overhead”. Sizes are the stored, compressed sizes — the amount each part contributes to the file you have.
From diagnosis to a smaller file
Once you know the cause, pick the matching fix. For image-heavy files, the PDF Compression Analyzer estimates how much each image would shrink at a chosen resolution and quality before you change anything. For scans and photo documents, Compress PDF re-saves every page as an optimised image, which reliably cuts size but makes text unselectable.
If the PDF carries attachments or pages you don’t need, removing them is lossless: delete pages with Delete PDF Pages and review attachments before sending.
PDF Size Analyzer FAQ
Why is my PDF file so big?
Usually because of images: scans and phone photos store far more pixels than a printed page needs, sometimes without JPEG compression. Fully embedded fonts, attachments and repeated incremental saves are the next most common reasons. The PDF Size Analyzer shows which of these applies to your file.
Is the PDF Size Analyzer free and private?
Yes. It is free with no sign-up, and the file is analysed entirely in your browser — nothing is uploaded.
What are “unused objects”?
Objects stored in the file that nothing in the document refers to, typically left behind after pages or images were deleted in an editor. They add size but are never displayed.
What does “Structure & overhead” include?
Page dictionaries, the page tree, annotations, links, form-field definitions, the cross-reference table and older versions of objects kept by incremental saves. These small objects are often compressed together, so they are reported as one total.
Why do the image sizes differ from the files I inserted?
PDF software often re-encodes images when it creates the PDF — for example converting a PNG to Flate-compressed pixels or downsampling a photo. The analyzer shows the size actually stored in the PDF.
Can it open password-protected PDFs?
Sizes are measured from the file structure, so files with an owner password usually work. If a file can’t be parsed at all, remove the password first and try again.