From—
ToWebP
Output—
Change—
Convert a picture to PNG, JPEG, or WebP in your browser — and see all three encodings at once, with their real file sizes, before you commit. The smallest option is marked, transparency handling is spelled out, and one click adopts whichever format wins on your own image.
Step by step
01Add the file
Anything your browser can decode is accepted as input — including HEIC-adjacent formats it already understands, AVIF, and SVG. The decode happens locally.
02Choose the destination format
PNG keeps every pixel and the alpha channel. JPEG is universal and lossy with no transparency. WebP is lossy, keeps alpha, and is usually a quarter to a third smaller than JPEG at the same look.
03Set the quality, and the background if it matters
The quality slider applies to JPEG and WebP; PNG ignores it because there is nothing to trade. The background colour is what transparency becomes when the destination has no alpha channel.
04Compare the sizes, then download
Every candidate format is encoded so you can see the real byte cost side by side rather than trusting a rule of thumb about which is smaller.
Worked example
The comparison the tool runs on every change. The winner depends on what is in the picture — which is exactly why it encodes all three instead of recommending one.
Given
What each destination does to the same pixels
PNG lossless, alpha kept
every glyph edge exact, largest file
WebP lossy, alpha kept
smallest of the three, slight ringing on text
JPEG lossy, alpha composited onto the background
ringing on text, white fill where alpha wasFor a photograph the order flips: WebP and JPEG both win by a wide margin and PNG becomes absurd, often several times larger than the lossy pair. The rule worth remembering is about content, not format — lossy codecs are built for continuous tone and hate hard edges, so photographs compress beautifully and screenshots of text do not.
Before you convert
Content decides the format, not the other way round
Continuous tone — photographs, gradients, anything the eye reads as smooth — belongs in a lossy codec. Hard edges, flat fills, and text belong in a lossless one. A screenshot saved as JPEG and a photograph saved as PNG are the two classic mistakes, and they are the same mistake twice.
Converting cannot undo a previous loss
A JPEG re-saved as PNG is a lossless copy of a lossy image: every artefact from the first encode is now preserved perfectly, in a much larger file. Conversion changes the container, never the damage already inside it.
Transparency does not survive JPEG
There is no alpha channel in the format, so transparent pixels are composited onto the background colour before encoding. Set that colour deliberately — the default white is a decision, not a neutral.
WebP is the default worth defaulting to
It keeps alpha like PNG, compresses continuous tone like JPEG, and is supported by every browser in current use. The remaining reasons to choose JPEG are old software, print workflows, and systems that validate on file extension.
Quality numbers are not comparable between codecs
JPEG 90 and WebP 90 are two different encoders' scales; they do not mean the same thing and will not produce the same file. Judge by the measured bytes and by looking, not by keeping the number constant across a switch.
Everything runs in the browser
Decoding and re-encoding both use the canvas on your machine, so unreleased artwork and client photography never leave it. The share link carries the format, quality, and background — never the picture.
The judgement call
By what is in the image, since that is what the codecs actually respond to.
A photograph for the web
WebP
A quarter to a third smaller than JPEG at the same look, and alpha survives if the shot has a cut-out.
A photograph for a system you do not control
JPEG
Still the format everything reads. Anything that validates by extension or predates 2020 is happier with it.
A screenshot, diagram, or UI capture
PNG
Text and hard edges are the worst case for lossy codecs. Lossless keeps every glyph exact, and flat colour compresses well anyway.
A logo with a transparent background
PNG or WebP
Both keep the alpha channel. Choosing JPEG here is what puts a white rectangle behind the mark.
An icon or a piece of line art you have as SVG
Stay vector
Converting to any raster format fixes it to one resolution. Ship the SVG unless something downstream refuses it.
Trying to rescue a heavily compressed JPEG
Nothing helps
Conversion preserves the artefacts. The only fix is a better source file.
Reference
FAQ
Drop the PNG in and choose JPEG in the Convert to section. Because JPEG has no alpha channel, set the matte colour to whatever the image will sit on — white for most documents — so transparent regions blend in instead of appearing as a hard block. Then check the JPEG card for the resulting file size and download.
WebP for photographs and most illustrations: it is lossy, supports transparency, and is normally the smallest of the three at equivalent quality. PNG for logos, screenshots, and anything with flat colour and sharp edges, where lossless encoding both looks better and often compresses smaller. JPEG when maximum compatibility with older software matters. The comparison cards settle it for your specific image.
WebP as written here is lossy, so yes — controlled by the quality slider. At 90% the difference is very hard to see on a photograph while the file is typically a third smaller than JPEG. Converting from a lossy source compounds the loss slightly, since the encoder also has to reproduce the previous codec's artefacts.
No. Width and height pass through unchanged and no resampling happens. If you also want the picture smaller in pixels, resize it first with the image resizer, or use the compressor's longest-edge control.
Both work as input — your browser decodes them and the tool re-encodes to PNG, JPEG, or WebP. For an animated GIF only the first frame is converted, because a canvas holds a single still image. Those three formats are the outputs every browser can reliably write.
No. Decoding and encoding both happen on your device through the browser's canvas API. Nothing is transmitted or stored, and the tool works with the network disconnected.
Conversion happens in your browser. The file is decoded, redrawn on a canvas, and re-encoded locally — nothing is uploaded, and share links carry settings only.
Keep going
Image compressor
Hit a file-size budget with a quality search and a before/after wipe.
Image resizer
Change pixel dimensions with cover, contain, or stretch fitting.
Base64 converter
Turn an image into a data URI for CSS, HTML, or JSON payloads.
Metadata viewer
Inspect the EXIF and GPS tags a converted file leaves behind.