1.3.0 Rounded Corners, Trim, and Color Control
Shape your images with rounded corners (yes, even Apple-style squircles), automatically trim borders, and take control of color output with chroma subsampling and colourspace options.
This release is all about giving you more control over how your images look. Whether you want smooth rounded corners, need to crop out boring borders, or want pixel-perfect color output - we’ve got you covered.
Rounded corners
You can now add rounded corners to any image. We’ve got three different modes depending on what you need:
Percentage mode - Set the corner radius as a percentage of the image. 10% gives you subtle rounding, 50% turns a square into a circle. This scales with your image, so it looks consistent regardless of dimensions.
Pixel mode - If you need exact pixel values, like “exactly 20px radius on all corners,” this is for you. Great when you need things to match a specific design.
Squircle mode - The fun one. Apple uses these for their app icons - it’s a superellipse that looks smoother than regular rounded rectangles. Slide from subtle (lower values) to nearly circular (higher values). Your app icons will look right at home on iOS.
Pick your corner style: percentage, pixels, or that smooth Apple squircle shape
The preview updates live as you adjust, so you can dial it in before converting. Pro tip: use PNG or WebP output to keep the transparent corners instead of having them filled with white.
Trim to visible pixels (Pro)
Ever had an image with a bunch of transparent pixels, white space or uniform color around the edges? The new Trim feature automatically crops it down to just the visible content.
It detects uniform borders - whether that’s white, transparent, or any solid color - and removes them. Super handy for:
- Product photos with too much background
- Screenshots with extra padding
- Scanned documents with margins
- Icons extracted from sprite sheets
Just check the “Trim” box in advanced options and we handle the rest. Works with batch uploads too, so you can clean up a whole folder of images at once.
Before and after: automatic border detection and removal
Chroma subsampling (Pro)
This one’s for the color nerds (we mean that as a compliment). Chroma subsampling controls how color information is stored in JPEG, AVIF, and JP2 files.
4:4:4 - Full color resolution. Every pixel keeps its complete color data. Larger files, but no color bleeding on hard edges. Use this for graphics, text, or anything with sharp color transitions.
4:2:2 - Horizontal color reduction. Good balance between size and quality.
4:2:0 - The standard for photos. Color info is reduced both horizontally and vertically. Smallest files, and honestly you won’t notice on photos.
Auto - Let us pick based on your quality setting. High quality gets 4:4:4, lower quality gets 4:2:0.
Most people should just leave it on Auto. But if you’re exporting a UI screenshot and notice weird color fringing around text - bump it up to 4:4:4.
Colourspace output (Pro)
You can now convert images to different colourspaces:
- sRGB - Standard web/screen color (default)
- RGB - Generic RGB
- CMYK - For print workflows
- Lab - Device-independent color
- B&W - True grayscale output
The grayscale option is properly converted at the colourspace level, not just desaturated. So you get true single-channel output that’s smaller and more accurate than just sliding saturation to zero.
CMYK is particularly useful if you’re preparing images for professional printing. Export as TIFF + CMYK + 300 DPI and your print shop will be happy.
Pick your output colourspace - great for print prep or true grayscale
All features in the API
Everything above works through the API too. Here’s what we added:
edits.roundedCorners.mode = "percent" | "pixels" | "squircle"
edits.roundedCorners.value = 0-500
trim = true
chroma_subsampling = "4:4:4" | "4:2:2" | "4:2:0" | "auto"
colourspace = "srgb" | "rgb" | "cmyk" | "lab" | "b-w"
Check the API docs for curl examples and the complete parameter list.
What’s next
We’re working on some fun stuff for the next release. Can’t say too much yet, but batch operations are getting some love.