What WebP gives a website
WebP was released by Google in 2010 as a web-first replacement for JPG, PNG and GIF. Its lossy mode is based on the intra-frame coding of the VP8 video codec: it predicts each block of pixels from its neighbours and stores only the difference, which is more efficient than JPG's fixed 8×8 blocks. At the same visual quality a photo is typically a quarter to a third smaller.
On a single image that saving looks modest. On a page it adds up: a product listing with 40 photos at 200 KB each sheds roughly 2 to 3 MB once converted. Smaller images reach the screen sooner, which improves Largest Contentful Paint, the loading metric search engines use, and cuts mobile data use for visitors.
Every current browser displays WebP, and most content systems accept it for upload, including WordPress since version 5.8 and the major shop platforms. For a site built in the last few years there is rarely a reason to keep serving plain JPG.
When to keep the JPG instead
WebP is a delivery format, not an archive format. Keep your originals as JPG (or better, as the camera originals) and convert copies for publishing. Stay with JPG when:
- The file goes to print, a photo lab or a client who will open it in older desktop software.
- It is an email attachment for people whose mail apps you do not know.
- You are going to edit the image again later. Every lossy save costs some detail, and editing tools handle JPG more consistently.
If someone sends a WebP back and cannot open it, WebP to JPG reverses the process.
How this converter handles your files
The image is decoded by your browser, drawn on a canvas at its original size and encoded as WebP using the browser's built-in encoder, in a background worker when available. Files stay on your device.
Both JPG and PNG are accepted. PNG input is useful for screenshots and graphics with transparency: the alpha channel is kept, so a cut-out product shot still sits cleanly on any background. JPG has no transparency, so JPG input gives an opaque WebP.
Pixel dimensions are unchanged. The WebP format itself is limited to 16383 px per side, so very large panoramas need to be reduced before conversion. Metadata such as EXIF camera details and GPS location is not copied into the output, which is usually what you want for images published online.
Quality settings that work
The slider runs from 10% to 100% and starts at 85%. WebP's quality scale is not the same as JPG's, so do not simply copy the number you used for JPG. Some reference points:
- 80–85% for photos on content pages. This is the range where WebP gives most of its saving with no visible change at normal viewing size.
- 70–75% for thumbnails, listing grids and background images where the image is small or partly covered by text.
- 90–95% for hero images on large screens and for screenshots with fine text. Lossy WebP stores colour at half resolution, so thin coloured lines and red text on a dark background can look slightly soft at lower settings.
Avoid 100% for photos. It produces a file that is often larger than the JPG you started from, without any visible gain, because the JPG's own compression artefacts are being preserved in detail.
Getting the most out of the conversion
A JPG that has already been heavily compressed will not shrink much further. Converting a 60 KB JPG saved at quality 60 may give only a few percent, and the second round of lossy compression can make blockiness more noticeable. The biggest gains come from high-quality originals.
The order of steps matters more than the exact quality value:
- Start from the largest, best-quality version you have.
- Resize to the size the page actually displays, for example 1600 px wide for a full-width image or 600 px for a card.
- Convert to WebP at 80–85% and compare the before and after sizes shown for each file.
Batches are supported, with a 50 MB limit per file. Every file in a batch uses the same quality, so if you mix photos and screenshots, convert them in two rounds with settings suited to each.