Compress PDF to 2 MB
For long documents. Two megabytes is generous per page — which is why splitting is sometimes the smarter fix.
Updated · Published by Novpix LLC
Drop a PDF here
or click to browse — target is 2 MB
Files are processed on your device and never uploaded.
Runs entirely in your browser — nothing is uploaded
How it works
- 1Drop your PDF onto the tool. This target suits long scans: contracts, deeds, medical records, thesis submissions.
- 2Click Compress. Pages are re-rendered at the highest resolution that keeps the whole document under 2 MB.
- 3Download the result — or, if the document runs to hundreds of pages, split it and compress the parts separately.
Frequently asked questions
- How long a document does 2 MB cover?
- Roughly sixty to eighty scanned pages at a readable resolution. Each imaged page costs about 20 to 40 kilobytes at minimum, so the ceiling is set by page count rather than by any quality setting. A two-hundred-page scan has a floor above 2 MB no matter what, and the tool will report that instead of returning an oversized file.
- When is splitting the document better than compressing it?
- Whenever the page count is what makes the file large. Compressing a three-hundred-page scan to fit 2 MB means dropping every page to a resolution where the text stops being reliably readable. Splitting it into three parts keeps each page sharp and each file well inside the limit — and most portals that cap at 2 MB accept multiple attachments.
- Does a lower resolution break OCR or searchability?
- It removes it entirely, whether or not OCR was involved. Re-rendering replaces the page contents with a picture, so any text layer — original or added by OCR — is gone from the output. For archives, records and anything that needs to be searched later, keep the original file and treat the compressed copy as a submission artefact.
- Why is my document taking so long to compress?
- Because your own machine is doing the rendering, once per page, several times over. A sixty-page document at two or three passes means well over a hundred page renders. Nothing is uploaded, so there is no server farm to hand the work to — the trade for privacy is your processor's time.
- The document is 2.1 MB. Is there a gentler option?
- Yes, and it is worth trying first. Removing blank pages and scanned separator sheets, or exporting again from the original application, often clears a small overshoot without rasterising anything. Re-rendering the entire document to save a hundred kilobytes is a poor trade when a few unnecessary pages would do it.
About this tool
Two megabytes is the limit that university portals, government gateways and records systems tend to set for full documents rather than single forms. It is a generous budget in per-page terms — around sixty to eighty scanned pages at a resolution where the text still reads properly — which changes the nature of the problem. At this target the question is rarely how hard to compress, but whether compression is the right move at all.
The ceiling comes from page count. Every page re-rendered as an image carries a floor cost of roughly 20 to 40 kilobytes before it stops being legible, so a document's minimum size is fixed by how many pages it has. When that floor sits above 2 MB, the tool reports the smallest size it achieved and the resolution it used rather than quietly handing back something over the limit.
That is exactly when splitting beats compressing. Squeezing a three-hundred-page scan into 2 MB means pushing every page to a resolution where small print becomes unreliable; splitting it into three files keeps each page sharp and each part comfortably inside the limit, and most systems that cap at 2 MB accept several attachments. If the overshoot is small, removing blank pages and scanner separator sheets often clears it without re-rendering anything.
Whichever route you take, remember what re-rendering costs on a long document: any text layer, including one added by OCR, disappears from the output. Keep the original for your records and searchability, and treat the compressed copy as a submission artefact. All of the work happens in your browser, which is why a long scan takes real time on your own processor — and why the document is never uploaded to anyone.