WebP Format

Convert Images to WebP: Real Benefits and Assumed Limits

Will converting all your images to WebP really improve your site? Not necessarily. And that's good news, it means we can choose smartly instead of migrating everything at once.

8 min read

Browser support in 2026

95.6% of actively-used browsers support WebP natively in 2026. Chrome has supported it since 2011. Firefox since 2019. Safari since iOS 14 and macOS Big Sur, both released in 2020. Edge since its switch to Chromium.

The Safari case long was the main argument against WebP migration. It doesn’t hold anymore. Since iOS 14, all iPhones and iPads display WebP correctly, no fallback necessary.

One real gray zone remains: email clients (Gmail, Outlook) generally don’t display WebP. And very old browsers (IE 11, Safari before 2020) don’t support it either. In practice, if your audience is mostly on recent devices and you don’t manage emails with embedded images, support isn’t an obstacle anymore.

Weight gains in practice

WebP compresses 25 to 34% better than JPEG at equivalent visual quality. This figure comes from Google’s comparative study on a corpus of thousands of real images, measured at equivalent SSIM index.

The range varies by content:

• Natural photos (landscapes, portraits, texture): gains near 27–34%
• Flat-color images (screenshots, UI, illustrations): gains toward 20-25%
• Already heavily-compressed JPEGs (quality 60-70): smaller gains, sometimes 10-15% only

A site with 50 average images of 300 KB in JPEG can drop to 200 KB average after WebP conversion. On LCP — the heaviest image in the viewport on load — this gain translates directly to milliseconds of loading. And LCP is a Google ranking signal since 2021.

Transparency and animation

WebP supports alpha channel (transparency) like PNG. A logo with transparent background converted from PNG to WebP lossless rarely loses visible quality. Weight gain averages 22 to 26% versus PNG per Google study, and can exceed 60% in lossy mode on simple logos with flat colors.

WebP also supports animations, like GIF. The gain here is spectacular: a 2 MB animated GIF converted to animated WebP typically yields 300 to 500 KB. Same logic as photos — WebP encodes successive frames much better.

The limits here aren’t technical but logistical. Some editing tools (old Photoshop versions, GIMP before 2.10, some legacy software) don’t handle WebP import. If your workflow involves frequent back-and-forth in an editing tool, WebP as a working format creates unnecessary friction. It remains primarily a web distribution format, not a production format.

When WebP isn’t the best choice

A fourth, subtler case: if your CDN handles WebP conversion on-the-fly (Cloudflare Images, Cloudinary, Imgix), converting images manually is double work. These services automatically serve the format based on browser capability: you upload JPEG, the user receives WebP if their browser supports it.

How to convert images to WebP

Three approaches depending on your context:

Online, no installation: Impmage converts your images to WebP directly in your browser, without sending files to a server. Squoosh (Google’s open-source tool) offers granular quality control and lets you visually compare before and after conversion.

Batch, locally: Sharp (Node.js library) and ImageMagick handle batch conversion via command line. Sharp especially is fast and reliable for processing entire folders. A typical ImageMagick command: magick mogrify -format webp -quality 82 *.jpg.

On-the-fly server-side: If your stack includes a CDN with image transformation, conversion happens automatically. You maintain one source file, the CDN serves the right format by browser.

Add JPEG fallback with picture tag

To cover browsers that don’t support WebP, the HTML <picture> tag lets you serve WebP to modern browsers and JPEG to others:

<picture>
  <source srcset="/images/photo.webp" type="image/webp" />
  <img src="/images/photo.jpg" alt="Precise image description" loading="lazy" />
</picture>

The browser loads the first source it supports. If WebP is recognized, it loads .webp. Otherwise, it falls back to .jpg. The loading="lazy" attribute applies to the <img> element and remains valid in both cases.

In practice, with 95.6% WebP support in 2026, this fallback concerns about 4.4% of traffic. The question is concrete: is this 4% in your audience? If your site targets businesses with legacy Windows 7 machines or populations with infrequently-renewed devices, the fallback remains relevant. For a general-audience blog on mobile, it’s almost pointless.

WebP vs AVIF: should we switch already?

AVIF has arrived. And technically, it outperforms WebP on nearly every compression metric. The question isn’t whether AVIF is better (it is) but whether its browser support justifies adopting it now.

The reasonable strategy today: WebP as primary format, JPEG fallback if your audience requires it. And monitor AVIF — in 12 to 18 months, its support will probably reach where WebP is today.

For a complete comparison of all available formats, the complete image formats guide for 2026 details WebP, AVIF, JPEG and PNG by each use case.

GlitchGhost

Independent Developer

Developer specializing in web performance and image optimization.

Web DeveloperPerformance SpecialistCreator of Impmage