Compression
How to compress an image to 100 KB without making text blurry
By ResizeFox · Published
A 100 KB image limit is a byte limit, not a resolution. Two pictures with identical width and height may need different compression settings because a plain background and a detailed street scene contain different information. Use this workflow when a form actually specifies a maximum file size; otherwise, there is no reason to force every picture under 100 KB.
Quick reference
| Use | Size or purpose | Note |
|---|---|---|
| Upload ceiling | 100 KB | Check whether the form means 100,000 or 102,400 bytes |
| Example source | 2400 × 1600 px | 3:2 photograph |
| Example smaller export | 1200 × 800 px | 3:2; one quarter as many pixels |
Step by step
- Read the destination's permitted formats, minimum dimensions and maximum bytes before editing. A smaller file is not useful if the form rejects its dimensions.
- Keep an untouched original. Crop only unused edges or background while preserving all required information.
- For a photograph, try JPG or WebP if the destination accepts it. For screenshots, text and line artwork, first compare PNG with WebP; JPG may add visible halos around characters.
- Use the ResizeFox 100 KB tool on a copy. Download the result only when the tool confirms the file actually meets the size target; do not assume a selected target guarantees success.
- Open the downloaded file at normal viewing size. Read the smallest text, inspect edges and verify the width and height. If it fails either the visual or dimension check, try another format or a less aggressive crop.
Why 100 KB cannot guarantee identical quality
Image encoding removes or represents information differently depending on the format and the amount of detail. Smooth gradients, hair, foliage and small type react differently to the same output budget. This is why no universal quality slider value will always produce a readable 100 KB file.
If the destination needs a minimum of 600 × 600 pixels, shrinking a 2000 × 2000 source to 300 × 300 may satisfy the byte limit but fail the submission. Check both constraints instead of optimizing only one number.
Worked geometry example: reduce pixels before lowering quality
A 2400 × 1600 image contains 3,840,000 pixels. A proportionally resized 1200 × 800 copy contains 960,000 pixels, exactly one quarter as many. That gives an encoder fewer samples to store, but does not mean the result will be one quarter of the bytes: format, image detail and metadata also matter.
For a photo that will only appear as a small web thumbnail, this trade-off may be reasonable. For a scan with small writing, reducing both dimensions by half may make characters impossible to read. In that case crop unnecessary borders first and use the largest readable size that fits.
What to do when the tool cannot reach the target
A file-size target is not a promise. Some combinations of image content, dimensions and supported export formats cannot reach 100 KB while preserving acceptable quality. If the tool reports that it cannot meet the limit, do not relabel or merely rename the downloaded file.
Check whether a smaller crop is allowed. If the form accepts multiple formats, compare JPG and WebP outputs; if it requires PNG, keep PNG even if the compressed file is larger. When a document authority specifies mandatory content, never remove information just to satisfy a size cap.
See the tradeoffs yourself in the Image Compression Lab: compare text, gradients and transparent graphics with downloadable results.
Sources and further reading
Platform specifications and application rules can change. Refer to the current destination instructions; the worked calculations and preparation choices in this tutorial are not a guarantee of acceptance.