Skip to content

How to Get a PDF Under 100KB for a Government Form

7 min read Updated June 6, 2026

A target-driven method for compressing a PDF under 100KB without making your scan illegible enough to get rejected.


You filled out the form, attached your PDF, and the portal rejected it: “File must be under 100KB.” You try again with a different export setting and now it is 480KB. The cap does not move, so the file has to.

The frustrating part is that most compress-PDF pages hand you one button and hope. When it works, great. When it does not, you have no idea why, and you start deleting pages at random. The fix is to figure out what is actually making the file heavy, then apply the right lever in the right order until you hit the target.

TL;DR: Color scans saved at high DPI are almost always the culprit. Convert to grayscale, downsample to 200 to 300 DPI, recompress, drop unneeded pages, and only split as a last resort. Stop the moment the text is still clearly readable.

First, find out what is making it big

A PDF can be heavy for a few different reasons, and the reason decides the fix. Open the file and look at what it actually contains.

If your PDF came out of a scanner or a phone scanning app, every page is an image, even if it looks like text to you. That image carries a resolution (DPI, or dots per inch) and a color depth. A single A4 page scanned in full color at 600 DPI is a large photo. Ten of them is a slideshow of large photos wrapped in a PDF. This is the most common cause by far.

If your PDF was exported from a word processor or a design tool, the weight is more likely embedded fonts, a high-resolution logo or header image, or photos you placed in the document. The text itself is tiny in file terms.

Here is the quick map from cause to fix.

What is bloating the fileThe fix
Full-color scan of black-and-white textConvert to grayscale, or to pure black and white where the form allows it
Scan saved at 400 to 600 DPIDownsample to 200 to 300 DPI (often less for plain text)
High-resolution photos placed in the documentRecompress and resize the images to the size they actually display at
Embedded fonts you do not needSubset the fonts, or export with standard fonts
Blank, duplicate, or instruction pagesRemove the pages you were not asked to submit
Still over the cap after all of the aboveSplit into the separate documents the portal expects

Most files are heavy for the first two reasons. If you only learn one thing here, learn that color and DPI are the big levers.

The order of operations to hit an exact target

Do these in sequence. After each step, check the file size. The moment you are under the cap with readable text, stop. There is no prize for going smaller, and over-compressing gets you rejected.

1. Convert color scans to grayscale or black and white

A scan of a black-on-white document does not need color information for every pixel. Color roughly triples the data compared to grayscale. If the form is text, signatures, and stamps, grayscale loses nothing you need and cuts the size hard.

If the form explicitly allows pure black and white (also called bitonal or 1-bit), that goes even further, because each pixel becomes a single on or off value. Use this only for clean text documents. A black-and-white conversion of a photo ID or a faint signature can lose detail, so check the result before you commit.

Watch the exception: if the portal requires a color copy of an ID or a photo, skip this step for that page and lean harder on DPI and recompression.

2. Downsample the scan resolution

DPI is how many dots make up each inch of the scanned page. Scanners often default to 300, 400, or 600 DPI. For a document you are reading on screen or printing once, 200 to 300 DPI is plenty. Many text scans stay perfectly legible at 150 DPI.

Halving the DPI roughly quarters the image data, because resolution counts in both directions. Going from 600 DPI to 300 DPI is a large saving on its own and usually invisible to the eye. Going from 300 to 150 saves more but starts to show on small print, so check the smallest text on the page before you accept it.

The trap here is downsampling too far. If letters look soft, broken, or surrounded by a halo, you have gone past what a reviewer will accept. Step back up a level.

3. Recompress the images

Once color and resolution are right, the images can still be compressed further by how they are stored inside the PDF. This is a quality-versus-size dial. A moderate setting removes data your eye will not notice. A harsh setting introduces blocky smudging around text edges, which is exactly the kind of thing that gets a scan bounced.

Aim for the highest quality that still fits under the cap, not the smallest possible file. You usually have headroom once grayscale and downsampling have done their work.

4. Remove pages you were not asked to submit

Scans often include a cover sheet, a blank back page, instructions, or a duplicate. Every page you drop is a full page of image weight gone. Read the form’s instructions for exactly which pages they want, and remove the rest. This step alone can take a four-page scan under the cap.

5. Split, only if it still will not fit

If you have converted to grayscale, downsampled to a sensible DPI, recompressed, and trimmed pages, and the file is still over the limit, the document may simply be too much content for one 100KB file. At that point, check whether the portal accepts separate uploads. Many ask for each document as its own file anyway. Splitting a combined scan into individual documents, each under the cap, is cleaner than crushing one file until it is unreadable.

How far is too far

The whole point of this is to pass review, not just to pass the file picker. A scan that is under 100KB but illegible gets rejected by a human later, and now you have lost days waiting.

Before you submit, zoom in on the smallest text, any handwritten signature, official stamps, and ID numbers. If you can read all of it cleanly, you are fine. If anything is mushy, fringed with color artifacts, or broken up, back off one step: raise the DPI, lift the compression quality, or keep grayscale instead of pure black and white. A 95KB file that a clerk can read beats an 60KB file that gets returned.

Doing this privately

Government forms, IDs, payslips, and signed contracts are exactly the documents you do not want to send to an unknown server just to shrink them. The good news is that none of these steps require uploading anything. The conversion, downsampling, and recompression can all run locally in your browser, so the file never leaves your device.

Local PDF compression tools are coming soon to pdf.hivly.net, built around this target-driven approach: pick your size cap, choose how aggressive to go, and keep the document readable. Free, no sign-up, and the file stays on your machine.

Until then, the order above is the method. Color and DPI first, recompression next, pages after that, splitting last, and stop the moment the text is still sharp and you are under the cap.

Try the pdf toolsMerge, split, compress, protect, unlock, sign and convert PDFs to and from images.

Frequently asked questions

Why does my PDF stay large even after I compress it?
Most generic compressors only restructure the file. The real weight is usually a full-color scan saved at high DPI. Convert it to grayscale and downsample the scan resolution, and the size drops far more than recompression alone.
Will compressing a PDF make it too blurry to accept?
It can if you push too far. Keep document scans at 200 to 300 DPI and keep text sharp. If letters look fuzzy or fringed with color, you have gone too low and the form may be rejected.
Is 100KB even possible for a multi-page scan?
For one or two pages of black text, yes. For several full-page color scans it is harder. Convert to grayscale first, downsample, then remove blank or duplicate pages before considering a split.
Do I have to upload my document to a website to compress it?
No. The work can happen entirely in your browser so the file never leaves your device, which matters for IDs, payslips, and signed forms.

Keep reading

Building something bigger?

Hivly is made by CodingEagles, a software studio that ships production web apps. If you have a real project, get in touch.

See what CodingEagles does →