How this site works
What happens to a file you drop here, and why the work does not run in your browser.
The file is converted, then dropped
Your file goes to our server, is decoded by libvips, redrawn at the size you asked for and re-encoded in the target format. It lives in the memory of that process for the seconds the work takes and is never written to disk. When the result has been sent back, the original is gone: there is no bucket, no queue and no database of submissions, so there is nothing to retain and nothing to leak.
Re-encoding also drops the metadata the camera wrote, including GPS coordinates, so converting a photo before publishing it strips the location it was taken at. The single tag we do read is the orientation flag, so that a photograph taken sideways comes out upright instead of on its side.
Why not in the browser
A browser canvas only decodes what the browser already understands — in practice PNG, JPEG, WebP and, on newer systems, AVIF. It will not open TIFF or camera RAW, it writes no AVIF or GIF at all, it keeps only the first frame of an animation, and it quietly discards the colour profile along the way. Which formats worked would then depend on which browser you happened to open, and that is a poor answer to give someone who just wants a file converted.
libvips on our own server has one answer for everyone: the same formats, the same quality and the same options no matter what you browse with. That is worth the round trip, and it is why the privacy page states plainly what happens to the file while it is with us.
One engine, many tools
Every converter here is a definition rather than its own little application: the formats it accepts, the options it offers, the specification behind them and the results it has to reproduce. A shared engine renders it. That is why the tools behave consistently, and why a converter with no passing test cases never reaches the build.
Format behaviour is transcribed from the MDN image format guides and the relevant W3C specifications, and the citation stays next to the claim on the page. Where a format cannot store something — PNG keeps transparency, JPEG does not — the tool says so rather than silently discarding it.
What is here today
5 published converters, with 15 reference cases behind them.