Benchmark methodology

We publish how we measure — not just who wins. Every comparison below is reproducible from the same inputs and settings.

What we measure

• File size (bytes) before/after, absolute bytes saved, % saved
• Output dimensions, format, and quality/preset used
• Encoding time (ms) and processing location (server vs local)
• Transparency preserved (yes/no), metadata stripped/kept
• Visual check: side-by-side at 100% + zoom, no invented scores

Standard presets

What we will not claim

• No “best compressor” badge without a defined metric, dataset, and date.
• No “lossless” unless the output is mathematically lossless.
• No “73% smaller” marketing numbers — only measured per-image results.
• No uptime, certification, or scale claims without evidence.

Try the presets yourself: Optimize images · All tools · API docs