An upload limit is a byte budget. Start with the receiving site's actual limit and accepted formats, then make a copy of your image that fits while preserving the detail you need. A web performance budget is a separate choice: an image can pass an upload limit and still be unnecessarily large for a page.
Work out your file-size budget
Suppose your image is 4.7 MB and a form allows 1 MB. Using decimal units, you need to reduce 4,700,000 bytes to at most 1,000,000 bytes: a reduction of about 78.7%. That is a target, not a prediction of what a compression setting will achieve.
Required reduction = (1 - limit / original size) × 100%. Use the same units for both sizes. The Data Size Converter distinguishes MB from MiB: 1 MB is 1,000,000 bytes, while 1 MiB is 1,048,576 bytes. If a receiving service labels its limit ambiguously, leave some headroom and test the upload.
Byte-budget worksheet
How much smaller must the image be?
Enter the original file size and the receiving site's limit. Nothing is uploaded or stored.
Minimum reduction needed
78.7%
Remove at least 3,700,000 bytes. This is a byte target, not a predicted quality setting.
For a decimal KB target in the Image Compressor, enter no more than 1,000 KB. Check the actual exported bytes and leave headroom when the destination's limit is ambiguous.
| Example limit | Decimal byte budget | Check before exporting |
|---|---|---|
| 100 kB | 100,000 bytes | Small text and sharp edges |
| 500 kB | 500,000 bytes | Faces, gradients, and fine texture |
| 1 MB | 1,000,000 bytes | Dimensions and accepted format |
These are example budgets, not recommended limits for every website. The finished file's byte size is the deciding measurement; quality 50 does not mean half the original file size.
For a stated decimal limit such as 100 kB (100,000 bytes), enter 100 in the Image Compressor's optional KB target. The tool aims below that byte count and reports whether the exported file actually meets it. Some images cannot reach a tight target at their current dimensions; resize first or choose another accepted format rather than assuming the target was met.
Here are five methods to try. Their impact depends on the source image and its destination.
1. Compress the image
Lossy compression can shrink a file by discarding detail. Start with a high-quality setting and lower it in small steps while checking gradients, text, faces, and sharp edges. Different encoders use quality values differently, so a number from one tool is not directly comparable with another.
Drop a copy into the Image Compressor to compare the before-and-after size and preview the result.
One important rule: lossy compression is one-way. Detail discarded by lossy JPG or WebP settings cannot be restored from that output. Lossless formats can preserve image detail, but re-encoding may still change file size and metadata. Always keep your originals and compress copies.
2. Convert to WebP
WebP supports lossy and lossless compression as well as transparency. Modern browsers support it, but the receiving app or publishing platform may still impose its own format rules.
If your platform accepts WebP, convert a copy with the PNG to WebP converter and compare both size and appearance with the original format.
3. Resize the dimensions first
Check dimensions before repeatedly lowering quality. A 4000×3000 photo has 12 million pixels. At an 800-pixel display width, an 800×600 version has 480,000 pixels; a 1600×1200 version is a 2× source for that display width, with twice the dimensions and four times the pixels. Choose based on display size, sharpness requirements, and the destination's rules.
Resize close to the largest dimensions the destination will display before compressing. Fewer pixels usually means less data to store, although the final file size will also depend on format and content. Use the Image Resizer to choose exact dimensions.
4. Strip metadata
Photos from phones and cameras can carry EXIF metadata such as location, camera settings, timestamps, and thumbnail previews. Removing metadata can reduce size and avoid sharing information the destination does not need.
Some image tools remove metadata when re-encoding. Do not assume every export strips EXIF: inspect the resulting file before sharing if location or other metadata matters.
5. Pick the right format
Not every image should be a JPEG. Here's the quick rule:
- Photos → JPEG or WebP
- Screenshots, graphics, text overlays → PNG or WebP
- Simple graphics, logos, icons → SVG
Saving a screenshot as JPEG can create artifacts around text and sharp edges. Saving a detailed photo as PNG can create a larger file than a suitable lossy format. Match the format to the content and compare the actual outputs.
If compatibility or transparency makes that choice less obvious, use the PNG, JPG, and WebP comparison before converting the file.
Putting it all together
Use a small comparison table for your own source file rather than relying on a generic benchmark:
| Step | Action | File size |
|---|---|---|
| Start | Original dimensions and format | Record size |
| 1 | Resize to destination dimensions | Record size |
| 2 | Compress a copy | Record size |
| 3 | Try an accepted alternate format | Record size |
Compare each version at its intended display size and keep the smallest one that still preserves the detail your audience needs. Resizing before the final compression pass also avoids encoding pixels that the destination will not show.
How much can I compress an image without losing quality?
There is no reliable universal percentage. Test a copy, compare it with the original at the intended display size, and keep the original available in case you need a different output later.
What's the best image format for small file sizes?
For photos, compare JPG and WebP when the destination accepts both. For screenshots and graphics with text, compare lossless PNG or WebP. If compatibility is uncertain, check the destination's accepted formats before converting.
Does resizing an image reduce file size?
Usually, but the byte savings depend on encoding and image detail. Resizing 4000×3000 to 1920×1440 reduces the pixel count from 12,000,000 to 2,764,800, a 76.96% reduction. The file-size reduction need not match that percentage. Check the exported file after resizing and compression.