How to Compress a PDF Under 2 MB Without Wrecking the Text
The short answer
To get a PDF under 2 MB without making the text blurry, shrink the pictures inside it, not the whole page. Photos and scans are almost always what makes a PDF heavy. Re-saving them at a sensible resolution (about 120 to 150 dots per inch of printed size) usually brings a scanned document under the limit while the typed text stays crisp, selectable and searchable.
Turning every page into a flat image is the last resort, not the first move. It throws away the text layer, and on text-heavy files it can even make the PDF bigger.
If you want to do it now, open Compress PDF, choose “Get it under a size”, pick 2 MB and press Compress. The rest of this guide explains what is happening, so you can fix the files that don’t cooperate.
What actually makes a PDF big
A PDF is a container. The ISO 32000 standard that defines the format describes a file as a collection of objects: pages, fonts, text drawing instructions, and “sampled images”, which is the spec’s name for photos and scans. Each of those costs bytes in a very different way.
| What’s inside | How much space it takes | Can it shrink safely? |
|---|---|---|
| Typed text (from Word, Google Docs, a bank) | Very little: a page of text is a few kilobytes of instructions | Not much. It is already compact. |
| Embedded fonts | Tens to hundreds of KB per font, once per file | Rarely worth touching |
| Vector drawings and charts | Small to moderate | Usually leave as is |
| Photos, scans, screenshots | Often 0.5 to 5 MB per page | Yes, by lowering resolution and re-saving as JPEG |
| Leftover junk from repeated edits | Varies | Yes, losslessly |
The numbers in that middle column are typical ranges, not rules, but the pattern holds: if a PDF is big, look at the pictures first.
Why phone scans are so heavy
Here is the arithmetic for a common case. A 12-megapixel phone camera produces a 4032 × 3024 pixel photo. Put one of those on a Letter page (8.5 × 11 inches) and the long side gives you 4032 ÷ 11 ≈ 367 pixels per inch.
That is far more detail than you need. A screen shows a page comfortably at around 100 to 150 pixels per inch, and 150 DPI still prints a readable page on an office printer.
Now compare pixel counts, because JPEG file size roughly follows the number of pixels:
| Target resolution | Pixels on a Letter page | Share of the original 12.2 MP |
|---|---|---|
| 367 ppi (as shot) | 3,024 × 4,032 ≈ 12.2 MP | 100% |
| 150 DPI | 1,275 × 1,650 ≈ 2.1 MP | about 17% |
| 120 DPI | 1,020 × 1,320 ≈ 1.3 MP | about 11% |
| 96 DPI | 816 × 1,056 ≈ 0.9 MP | about 7% |
Dropping to 120 DPI keeps roughly one ninth of the pixels. So if a 10-page phone-scanned lease weighs 20 MB, about 2 MB per page, a rough estimate is 2 to 3 MB at 120 DPI and around 1.5 MB at 96 DPI, with the text on the scan still easy to read. How much you save on your own file depends on how it was made, so treat this as an estimate and check the result.
Why a text-only PDF barely moves
A 40-page contract exported from Word might be 600 KB. Almost all of that is text instructions and a couple of embedded fonts. There are no pictures to resample, so a compressor can only do lossless housekeeping, and you’ll see a few percent at most.
That housekeeping is still useful on files from older software. Since PDF 1.5, objects can be packed into compressed “object streams”, and tools like qpdf can regenerate those and recompress data streams without changing how anything looks. On a file from a careless generator that can be a real saving; on a clean export it’s small.
“2 MB” is not always the same number
Upload forms are not consistent about what a megabyte is. NIST spells out the two meanings:
- 1 MB (decimal) = 1,000,000 bytes, so 2 MB = 2,000,000 bytes.
- 1 MiB (binary) = 1,048,576 bytes, so 2 MiB = 2,097,152 bytes.
Plenty of software still writes “MB” when it means the binary unit. Windows File Explorer, for example, divides by 1,024, so a file of exactly 2,000,000 bytes shows up as about 1.90 MB there.
The safe move is to aim under 2,000,000 bytes. A file that size passes a portal using either definition. Our compressor targets the decimal number for exactly this reason.
Step by step: get a PDF under 2 MB
- Open the file in Compress PDF. It runs in your browser, so the file isn’t uploaded. That matters for leases, tax forms and medical records.
- Choose “Get it under a size” and pick your limit. The presets are 500 KB, 1 MB, 2 MB, 3 MB, 5 MB, 10 MB and 25 MB, or type a custom size.
- Press Compress. The tool works through a ladder of settings from gentle (200 DPI, high JPEG quality) to harsh (50 DPI). It checks the harshest first, so an impossible target is spotted right away, then settles on the gentlest setting that fits.
- Read the result line. It shows the before and after size and confirms the text is still selectable. If the result wouldn’t be smaller, you keep your original.
- Open the new file and zoom to 100%. Check small print, signatures and any numbers someone will need to read, like account or case numbers.
If you don’t have a hard limit and just want a smaller file, use “Make it smaller” instead. Light (150 DPI) suits anything that may be printed. Smallest (96 DPI) is fine for documents that will only be read on screen.
When it still won’t fit
Sometimes even the harshest setting misses. That tells you something about the file.
The PDF has lots of pages of pure text. No picture is left to shrink. Your options are to split it, or to ask whether the portal accepts several files. Split PDF can cut it by page ranges or every N pages.
Some pages don’t need to be there. Blank backs of double-sided scans, cover sheets and duplicate pages all cost space. Remove them with Delete PDF pages before compressing.
The scan itself was made badly. If you control the scanner, rescan at 200 or 300 DPI in grayscale instead of color photo mode. A black-and-white text page doesn’t need color, and color costs bytes.
You really need it smaller and don’t need the text. Our tool offers “flatten pages to pictures” only after the gentle method fails, and warns you first. Use it knowing that text stops being selectable or searchable, links and form fields stop working, and screen readers can’t read the result.
Common mistakes
- Flattening first. Turning pages into images is the reflex of many tools, but it destroys search, copy and accessibility. On a text PDF it often makes the file bigger, because a picture of a page is heavier than the instructions to draw it.
- Zipping the PDF. Most portals won’t accept a ZIP, and zipping barely shrinks a PDF whose images are already JPEG-compressed.
- Compressing the same file over and over. Each lossy pass softens images a bit more. Go back to the original and choose a stronger setting once.
- Taking screenshots of pages. You lose the text layer and usually end up with a heavier file.
- Forgetting about email. Personal Gmail accounts can send up to 25 MB of attachments. Over that, Gmail swaps the attachment for a Google Drive link, which some recipients can’t open. The 25 MB preset is there for that case.
- Not checking the output. A compressor can’t see whether a faint pencil note in the margin matters to you. Zoom in before you submit.
A quick decision guide
| Your situation | What to do |
|---|---|
| Phone-scanned or photo-heavy PDF over the limit | Compress with “Get it under a size” |
| Text-only PDF over the limit | Split it, or remove pages you don’t need |
| Mixed PDF just over the limit | Compress at Balanced (120 DPI) |
| Needs to print well later | Compress at Light (150 DPI) |
| Portal limit is tiny (under 500 KB) and you have many pages | Ask for a split upload; flattening is the last resort |
The general rule: shrink pictures, keep text, and never send a flattened file when a gently compressed one fits.