How to make image files smaller (without making them ugly)
Reviewed July 28, 2026
Boltpic
Each guide describes the current Boltpic tools and links to relevant technical references. Results that depend on the image, browser, or device are identified as variable.
Upload forms set different limits and accept different formats. A phone original is often far larger than the size at which it will be displayed, so sending it unchanged can fail the limit.
The controls you can act on are pixel dimensions, format, and compression quality. Adjust them in that order and inspect each export to preserve the detail the destination actually needs.
Lever one: dimensions
Halve both width and height and the total pixel count falls to a quarter. If the source is larger than its display size, matching the dimensions before lowering quality usually removes more weight while protecting detail.
For example, an image displayed at 800 pixels wide can use a 1,600-pixel candidate for high-density screens. If the source is much larger, the extra pixels may only add transfer weight. A zoom feature or original-file download is a different requirement and may justify a larger candidate.
Measure the rendered size instead of memorizing dimensions. Prepare a normal candidate and a 2× candidate for high-density screens, then use srcset and sizes to let the browser choose.
Lever two: quality (push until just before it shows)
For JPG and lossy WebP, the quality setting controls how much detail is discarded. The same number can produce different results across browsers and encoders, so it is not an absolute scale. As you lower it, inspect smooth areas such as skies, skin, and out-of-focus backgrounds for banding and block artifacts.
Try a quality around 80 as a starting point, compare the result beside the original, and go lower only while the result remains fit for its purpose. Boltpic's compressor shows both images and file sizes together so you can judge the pixels and the saving at the same time.
Quality 100 does not guarantee the original pixels. Re-encoding to a lossy format can still change the image and may create a needlessly large delivery file. If preserving the original matters, keep the source file separately.
Lever three: format
Once dimensions are right, compare formats. For a web photograph, export both WebP and JPG and inspect the smaller result. PNG often suits screenshots with text and graphics that need transparency, while JPG remains a practical choice for broad compatibility.
Four ways people waste quality
- Compressing twice. Every lossy save stacks new artifacts on old ones. Keep one original, compress once, and re-do it from the original if you need a different setting.
- Enlarging small images without checking the result. Upscaling cannot restore missing detail and may look soft; enlarge only when the destination requires it.
- Saving text-heavy images as JPG. Text edges are exactly what JPG handles worst. Screenshots stay PNG.
- Trusting the quality number across tools. One tool's “80” is another's “90.” Judge with your eyes, not the number.
Putting it together: meeting an upload limit
Start with the destination's recommended dimensions and accepted formats. If the source is larger than its display size, resize it first. Export a photograph as WebP or JPG, then lower quality in small steps until it meets the limit. Because every image compresses differently, check both the downloaded file size and fine detail at a useful zoom.
Boltpic's resize and compress tools handle this workflow in the browser. Keep the source, compare a few exports, and use only the result that matches the destination.