Canvas Size
Set a new canvas around a JPEG, PNG, WebP, AVIF, GIF, BMP, ICO, or HEIC image, in pixels or as a percentage of that picture. The picture stays the same pixel size and sits on the anchor you pick.
About this tool
Resize Image scales the picture. Canvas Size does not. It changes only the canvas around the existing pixels. The anchor grid chooses which point of the image stays fixed: a corner, an edge, or the center. Expanding adds space on the other sides. Shrinking keeps the anchored portion and drops the rest, still without scaling.
The lock keeps the canvas shape while you edit one side. It does not resize the picture. Leave it off when the new canvas should have its own width and height.
Percentage mode uses a single scale on the open picture. A setting of 120% turns 1000 by 800 into a 1200 by 960 canvas and turns 640 by 480 into 768 by 576. The artwork stays at its original pixel size, and the anchor places it. That is the setting to record when a batch contains photos of different sizes. The line under the scale field shows the pixel canvas for the file that is loaded. Dimensions mode is unchanged: you type the canvas in pixels, and the lock only ties those two fields together.
New areas use transparency when the result can store alpha, including PNG, WebP, AVIF, GIF, and ICO. A HEIC edit is written as PNG, so transparency can remain. JPEG and BMP cannot keep alpha, so those areas are painted with the fill color and the page says so. GIF keeps one-bit transparency, not partial alpha.
The preview is the canvas you will download. A checkerboard behind the image shows transparent pixels. The download and the next-tool handoff use that same image, not a separate sketch.
How to use it
- Choose an image. Width and height start at the current image size.
- Stay on Dimensions to type pixels. Switch to Percentage to size the canvas from the picture, such as 120%.
- Pick a point on the 3×3 anchor, choose transparent or a fill color, then apply the canvas.
Limitations
- The image is placed at its original pixel size. A smaller canvas clips whatever falls outside the anchor. Percentage mode rounds each side. This is not the crop tool.
- Transparent is offered when the output can keep alpha. JPEG and BMP use the fill color instead. GIF transparency is one bit. Files over 250 MB or 40 million pixels are refused.
Padding or cropping the frame without scaling
The picture stays its own size
Canvas size changes the frame, not the scale of the artwork. The decoded bitmap is drawn one-to-one, with smoothing turned off, at an anchor. Nine anchors exist: the corners, the edge midpoints, and the center. An unrecognized anchor value falls back to center. The top-left of the original lands at a computed offset, so a larger canvas leaves a margin and a smaller canvas clips whatever falls outside the new edges. Nothing in this path stretches the photo to fill the new width. Use the resizer when the picture itself should change pixel size. The optional aspect lock, off when the page loads, ties the two canvas fields together. It does not scale the image.
Percent mode multiplies that frame by one scale and still draws the bitmap one-to-one. Enter 150 and an 800 by 500 photo asks for a 1200 by 750 frame, while a 400 by 300 photo asks for 600 by 450. Each side is rounded on its own. A scale under 100 clips from the anchor you picked. The width and height boxes are skipped, so one saved step pads every file in a batch from that file’s own size.
New pixels come from the fill. Transparent fill leaves the canvas clear before the image is drawn, which PNG, WebP, AVIF, and a HEIC result written as PNG can store. JPEG and BMP cannot keep an alpha channel, so the transparent choice is hidden for those sources and a solid color is painted instead. The note on the result says the empty areas used that color. GIF can keep only a single fully transparent palette index, so a soft fade in the margin will not survive as a fade.
Encoding the larger or smaller frame
Preserve intent writes the source MIME when it can. The quality range is preset to 0.92 and is active for JPEG and for WebP. AVIF keeps it active only after two AVIF encodes of a probe image come out with different lengths. Slider stays disabled for PNG padding, for GIF padding, for a BMP pad, for an ICO pad, and for HEIC padding stored as PNG. If AVIF encoding is missing in this browser, the preserve path substitutes PNG and the pixels, including any transparent margin, are stored there.
A JPEG pad is a new DCT file. Camera EXIF from the original is not in that file. A PNG pad is lossless for the samples that remain, including a transparent margin, but it is still a newly compressed PNG rather than the original chunks with a larger IHDR.
Repeated saves on a padded JPEG
Each extra JPEG or WebP or AVIF generation can soften the original photo even though you only added a border of flat color. Flat color is cheap for those codecs. The photograph in the middle is not. Decide the quality when you first expand the canvas if that JPEG is the deliverable. Expanding a PNG, then exporting JPEG on a different page, moves the loss to that later export, which is often what you want when the pad must stay editable.
Privacy
Your file is processed in your browser and is not uploaded to UtilChain. The file name is not sent to analytics.

