JPG
WEBP
WEBP
ЗАГРУЗИТЬ
ФАЙЛ
Ожидание файла...
Превью
Ожидание файла...
Tool guide
Feed it a JPEG - a camera shot, a stock download, a screenshot someone re-saved - and get the same frame back as WebP, usually a quarter to a third smaller at comparable visible quality. Encoding runs on the Canvas API inside your own browser, so the photo is never uploaded anywhere. This is the step people take when a product gallery or article hero is dragging Largest Contentful Paint down and the layout itself cannot change. The file picker accepts image/jpeg only. See also: turn a transparent PNG into WebP, squeeze a photo down to a target size in KB, open an AVIF file and save it as JPG.
For a normal photograph, expect 25-40% off the JPEG size. A picture that has already been through a messenger or a CMS resize has lost detail before it reached you, and the gain can drop to 5-10%. Check the resulting size: if the saving is marginal, keep the JPEG rather than shipping two formats.
The file input is set to accept image/jpeg, so this page is scoped to JPG and JPEG. PNG sources have their own PNG to WebP page, and iPhone photos go through the HEIC converter first. A file named .jpg that is really a PNG inside will also fail to decode and no preview will appear.
Chrome, Firefox, Edge, Opera and Safari 14 and newer (macOS Big Sur, iOS 14) all render WebP. Old Safari builds and desktop mail clients such as Outlook do not, so avoid WebP in email. On the web, a picture element with a WebP source and a JPEG fallback covers everything.
Yes. This is lossy WebP applied on top of the loss already baked into the JPEG. Photographs survive it invisibly; screenshots with small text and charts with thin lines can pick up soft edges. Keep that kind of image as PNG instead of pushing it through WebP.
No - the page is one file at a time: load, DOWNLOAD, RESET, repeat. For hundreds of images, cwebp on the command line or a bundler plugin is faster. This tool is for the handful of pictures that go with a single post, without installing anything.
No. The frame is redrawn through Canvas and metadata is not carried over, so camera model, capture date, GPS coordinates and copyright fields all disappear. That doubles as a privacy step before publishing, but archive the original JPEG if you need the EXIF later.
Nowhere. The file is read with FileReader, drawn into a Canvas and encoded by the browser itself; the page issues no network request carrying the image. It keeps working with the connection off once the tab is open, and speed depends only on your CPU.
Before: cover-1920x1080.jpg - 842 KB, straight off a phone
After: cover-1920x1080.webp - 386 KB at the same pixel dimensions
Before: sneakers-front.jpg - 1.4 MB, 2000x2000 px
After: sneakers-front.webp - 610 KB, 2000x2000 px, zoom detail intact
Your rating and feedback help decide what to improve next.