Short answer: choose a photo, set a target size in KB (or a quality level) and compress. Everything runs inside this tab — the image is never uploaded — and the result panel shows the exact bytes and dimensions that were produced.
How compression works here
- Add one image. The tool checks the file signature, not just the extension.
- Optionally type a target in KB and pick a maximum dimension.
- Choose the output format. Keeping the input format usually gives the best result.
- Press Compress image. The browser re-encodes it and measures the real output size.
- Download the file. The reported size is the byte count of the file you receive.
What it can and cannot do
- Best effort, not a guarantee. A 3 MB photograph can usually reach tens of kilobytes. A detailed PNG may not reach a very small target without visible loss, and the tool will say so.
- PNG is lossless. Quality steps do nothing for PNG, so the engine relies on downscaling. For photos, JPEG or WebP is far more efficient.
- No PDF, HEIC or SVG input. Only the three formats the browser can decode and re-encode reliably are accepted. SVG is excluded on purpose: it can carry script.
- Metadata is dropped. Colour profiles and EXIF are not preserved by canvas encoding. Orientation is applied to the pixels before encoding.