When you need the smallest possible JPEG — a thumbnail, a preview, a file under a strict upload cap — this tool compresses more aggressively at a lower quality. Expect a much smaller file with a slight, usually acceptable, softening. For higher fidelity, use JPEG optimisation instead.
What the quality dial actually trades away
Aggressive compression works by turning the JPEG quality factor down hard — here to a low setting — so the encoder quantises image detail more coarsely and writes far fewer bytes. The visible cost is concentrated in high-frequency areas: fine textures, sharp edges, and busy backgrounds soften first, while large smooth regions like sky or skin hold up well. That is exactly the right bargain for the jobs this tool is built for. A 120-pixel thumbnail, an avatar, a preview tile, or an image squeezing under a strict upload cap doesn't need pixel-perfect texture — it needs to be small and fast. When fidelity matters more than bytes, step back up to JPEG optimisation, which keeps quality high and shrinks files through encoding tricks instead of by discarding detail.
Picking the right tool for the byte budget
Think of compression strength as a spectrum and choose by use case rather than habit:
- Thumbnails and previews: aggressive compression, where softening is invisible at small sizes.
- Hero and content images: optimisation, which stays near-visually-lossless.
- Tight upload limits: aggressive compression first, then resize if it's still too big.
- Modern delivery: WebP for the smallest file of all.
Matching the strength to the slot stops you from shipping a blurry banner or a needlessly heavy thumbnail. The same source can go through more than one path if you need several sizes.
Stack compression with resizing for the hardest limits
When a single pass still leaves you over a cap, the two levers compound. Aggressive compression cuts bytes at the current dimensions; the resizer cuts dimensions, and because file size tracks pixel area, shrinking the longest edge often does more than the quality dial alone. The dependable order is resize the source to its real display size, then compress hard — that way the encoder isn't spending effort on pixels you'll never show. For a content pipeline, this pairing reliably forces oversized originals under email and form limits without manual fiddling. All of it sits one upload apart on jpegoptim, so iterating to a target size is quick.