How to get a PDF small enough to email (and what actually works)
To get a PDF under an email attachment limit with OnCloudPDF, first try Compress PDF, then remove or split pages with Delete Pages or Split PDF. Compress only optimizes the file structure and never recompresses images, so a PDF full of photos or scans may not shrink at all; for that you need a tool that downsamples images.
Updated · Runs in your browser, no upload, no account
Start by finding out why the PDF is big
A PDF is a container. Text and vector drawings take very little space; photographs and scanned pages are what make a file heavy. A 40-page report made of typed text is usually small, while a 10-page document made of phone photos or scanner output can be tens of megabytes because every page is a full-size picture.
That distinction decides which fix can work. Open the file in PDF Info to see its page count and size, then look at it in your viewer: if you can select the text with the cursor, the pages are mostly text; if you cannot, they are pictures. For picture-based PDFs, restructuring the file saves very little, and the only real savings come from removing pages or lowering image quality.
Email providers also differ in what they accept as an attachment, and the limit includes encoding overhead, so a file close to the limit can still bounce. Check your provider's current limit instead of relying on a number you remember, and aim to leave some margin.
What Compress PDF does and does not do
Compress PDF loads your file and saves it again with PDF object streams enabled. That packs the internal bookkeeping of the document more tightly. It does not touch the pages themselves: images are not re-encoded, resampled or downsampled, and there is no quality slider and no way to aim for a target size.
The result panel shows the original and the new size. If the new file is not smaller, the page tells you the file was already optimized instead of claiming a reduction. Files exported from word processors often shrink a little; PDFs already saved with object streams, and anything dominated by images, often barely change. Always compare the two sizes before you replace your original.
Because nothing visible changes, there is also no quality loss. That makes Compress a safe first attempt: it costs a few seconds and cannot damage how the document looks.
Options that really reduce size on OnCloudPDF
When Compress does not help, the reliable levers are sending fewer pages or sending the document in parts. Each one is exact and predictable, because pages are copied unchanged.
- Remove pages nobody needs. Delete Pages lets you click the thumbnails of cover sheets, blank pages or duplicates and download a new PDF without them. File size usually falls when the removed pages contained images, but the tool does not promise a particular reduction.
- Send only the pages that matter. Extract Pages keeps the pages you click and produces one new PDF, in ascending page order.
- Split into parts under the limit. Split PDF cuts one PDF into several by ranges or every N pages and returns a ZIP with one PDF per range. Pages are copied unchanged, so the parts add up to roughly the size of the original. Attach the parts to separate emails, or send a download link instead.
- Merge only what you need. If you are assembling a pack from several files, remove unneeded pages from each file before using Merge PDF; merging does not compress, so the output is roughly the sum of the inputs.
Why Grayscale and PDF to Scan can make files bigger
It is tempting to try Grayscale PDF on the theory that fewer colours means a smaller file. The code does not work that way. Grayscale renders every page to a picture at twice the page size, converts the pixels to grey, and saves each page as a JPEG at quality 0.92 inside a new PDF. Text and vector graphics become pixels.
The consequences are concrete. A PDF made of photographs can come out smaller, because one JPEG per page replaces the original images. A text-heavy PDF can come out similar or larger, because tiny text-and-vector pages are replaced by full-page images. In every case the text is no longer selectable or searchable, and links, form fields, bookmarks and annotations are not carried over. Use it when you want a grey, print-friendly copy, not as a compression tool.
PDF to Scan works the same way, rendering each page to a JPEG (quality 0.85) after adding light grain and a paper tint. It is meant to make a document look scanned, not to shrink it, and it does not add a text layer. Do not use either tool to prepare an email attachment unless you have checked the result size and you are happy to lose the text layer.
A worked example: a 12-page application pack
Imagine a job application pack: a cover letter, a CV, a certificate scanned on two pages and a portfolio sample, merged into one PDF that your email provider rejects. A sensible order of attempts looks like this.
- Open the pack in PDF Info and note the size and page count.
- Run Compress PDF. If the panel shows a useful reduction, download it and check the size again. If it says the file was already optimized, move on.
- Use Delete Pages to drop a blank separator page or a duplicate certificate scan, then compare the size.
- If the pack is still too big, use Split PDF to make two parts, for example the letter and CV in one file and the scans in another, and send them as two attachments.
- If the recipient must receive a single small file, reduce the images with a tool that supports image recompression (see below) and then verify that the scans remain legible.
Quick comparison of the options
| Option | What it does | Typical effect on size | Trade-off |
|---|---|---|---|
| Compress PDF | Re-saves the file with object streams | Small for text files, often none for image-heavy ones | No visible change; cannot recompress images |
| Delete Pages | Removes the pages you click | Larger when the removed pages hold images | Pages are gone; the original stays untouched |
| Extract Pages | Keeps only the pages you click | Depends on the pages kept | Selected pages stay in ascending order |
| Split PDF | Cuts into several PDFs in a ZIP | Total stays about the same; each part is smaller | Several attachments instead of one |
| Grayscale PDF | Renders each page to a grey JPEG | Can go down or up | Text becomes an image; links and forms are lost |
| PDF to Scan | Renders pages to a noisy JPEG | Not designed to shrink | Text becomes an image; fixed effect |
When to use a different tool
OnCloudPDF does not include image recompression. If you must keep every page and still reach a size limit, you need software that downsamples images, such as a desktop PDF editor or a server-based compressor. Keep two points in mind. First, a server-based compressor uploads your file, so read its retention policy before sending anything confidential. Second, check the output at full zoom: aggressive settings can make small print or signatures unreadable.
Another approach avoids the problem entirely: put the file in a cloud drive and email the link. For a comparison of approaches, see Free PDF tools compared.
Frequently asked questions
Why did Compress PDF barely change my file size?
Compress only re-saves the PDF with an optimized structure. It does not recompress images, so PDFs made mostly of photos or scans often shrink very little or not at all. The page tells you when the file was already optimized.
Can I make a scanned PDF smaller on OnCloudPDF?
Not by recompressing its images; the site has no such tool. You can delete or extract pages, or split the file into parts. For a real size reduction of scans, use software that downsamples images and check that the text stays legible.
Does converting a PDF to grayscale reduce its size?
Sometimes, but not reliably. Grayscale rasterizes each page into a JPEG, which can be smaller for photo-based PDFs and similar or larger for text-based ones. It also turns text into an image that cannot be selected or searched.
Is my file uploaded when I use these tools?
No. Compress, Delete Pages, Extract Pages, Split, Grayscale and PDF to Scan all process the file in your browser, so it is not uploaded to OnCloudPDF.