What does compressing a PDF to a target size do?
Compressing a PDF to a target size means reducing the file until it lands under a number you were given — 500 KB for a form, 1 MB for an upload portal, 2 MB for an email. This tool takes that number and works through a ladder of compression steps, gentlest first, stopping at the first one that fits. It then reports which step was applied and whether the target was actually reached, rather than claiming success either way.
A PDF has no single quality setting, so a target cannot be met by turning one dial. What occupies the space is almost always the photographs embedded in the file: they are large, they are lossy data, and they can be re-encoded at a lower quality or resolution. Text, fonts and vector art are already compact and change very little under any treatment. That is why a dense scanned document shrinks dramatically and an all-text report barely moves, and why a target that is reasonable for one is impossible for the other.
The ladder reflects that. It begins with a structural rebuild that re-encodes nothing and is therefore lossless. If that is not enough it steps down the image quality and resolution together, because lowering a resolution also lowers size and sometimes looks better than a low quality at full size. Only if nothing else reaches the target does it rasterise the pages, which reliably fixes the size but costs selectable text. The order is deliberate: the document should be damaged only as much as the target actually requires.
How to compress a PDF to a size
- Choose the PDF — up to 100 MB, since the compressor runs in the browser and holds the whole file in memory.
- Enter the target in kilobytes; multiply megabytes by 1024, so a 1 MB limit is 1024 KB.
- Press Compress to the target and watch the attempts reported as they run.
- Read the result: either it fits, with the size it reached, or it does not, with the smallest size that could be produced.
- Check what was applied — a rasterised result has lost its selectable text, so it is worth trying a larger target first.
- Download the file, or raise the target and run again if you would rather keep the text searchable.
Compress PDF to a Target Size limits and rules
| Item | Limit |
|---|---|
| Input | A single PDF, up to 100 MB |
| Target | A size in kilobytes you choose, from a few hundred upward |
| First step | A lossless structural rebuild, attempted only when the target is close to the original size |
| Then | Image re-encoding at 70%, 45% and 30% quality, with the resolution dropping alongside |
| Last step | Rasterisation, which fits almost any target but makes the text unselectable |
| Attempts | At most five, stopping at the first result under the target |
| Not covered | Removing pages, splitting the document, OCR, and any target below the floor a given document can reach |
When to use it
- Bringing a scanned statement under a portal's 2 MB upload limit.
- Preparing a brochure to travel as an email attachment.
- Meeting a form's size rule without hunting for the right quality setting.
- Reducing a photo-heavy report before sending it to a printer.
- Getting a document under a limit you have been given but not told how to meet.
Is it private?
The document is read and re-encoded in this page by a WebAssembly compressor. Nothing is uploaded, stored or logged, and there is no queue: closing the tab discards the file.
Frequently asked questions
Because every document has a floor, and below it the only way down is to remove content. A full page of scanned text is a large image, and no quality setting makes it small without making it unreadable — the rasterisation step will fit almost any target precisely because it throws information away. When even that is not enough, the number is genuinely out of reach for this document, and saying so is more useful than producing a file that does not open well.
Because rasterising is destructive in a way that is easy to overlook. Each page becomes an image: the text stops being selectable, stops being searchable, stops being readable by a screen reader, and will not scale cleanly when printed. For a document that is only ever viewed as a picture that is a fair trade, but it is the wrong default for a contract, an invoice or anything that will be searched later. Working up to it means it is used only when it is the only thing that works.
Both, and that is exactly the trap. Storage has traditionally counted in powers of two, so 1024 KB is a megabyte; but drive manufacturers and some cloud services count 1000 KB to the megabyte, and 'MB' in a user interface may mean either. The difference is about 2.4%, which is enough to fail a strict limit. If a form says 2 MB and you are close to the edge, aim for 1900 KB and remove the ambiguity.
No. The image step touches only the images: text, fonts and vector graphics are copied through and stay sharp, which is one of the reasons this approach is preferred over rasterisation. What can change is a photograph behind text or a scanned page, since both are images. That is also why a document that is all text will barely shrink — there is no image data to reduce, and the bytes left are the structure and the fonts, which are already compact.
The first one that fits, and if none fits, the smallest produced. The ladder stops as soon as it succeeds, so a document that fits at the gentle image step is never subjected to the harsher one — each step is a fresh compression of the original rather than a second pass over an already-compressed file. That matters because re-compressing an already compressed PDF compounds the damage, and this way it never happens.
Not from the compressed file, no — the discarded information is gone, which is why the original should be kept. If the result looks worse than you would like, raise the target and run the compression again from the original; each attempt starts fresh, so nothing carries over. Keeping the untouched original and producing a compressed copy for the specific upload is the habit that avoids every problem here.
Related tools
Written and maintained by Md. Nur Hossain Repon.