Bring Your Own Images: Drag, Drop, and Edit on the Canvas
Until today, anything on the Icon Maker canvas had to come from our generators. That's changed. You can now drag your own images straight onto the canvas, remove their backgrounds in one click, and use them as a base for AI edits.
Related: Editor V2, Icon Format Performance, Open the Editor
Three ways to add an image
Whatever's quickest for the moment: drop a file from your desktop, paste from the clipboard, or pick through a file dialog. They all end up the same once the image lands on the canvas.
Drag and drop
Drag a PNG, JPEG, or WebP from your desktop or browser straight onto the canvas. Up to five at once; the rest queue automatically.
Paste from clipboard
Cmd+V (or Ctrl+V on Windows) drops a copied image at the center of the canvas. Works from screenshots, browser image-copy, anywhere your clipboard can hold an image.
Add image button
The new "Add image" button in the layers panel opens a file picker. Multi-select works, 10 MB cap per file.
Uploaded images become regular layers. Select, move, resize, transform, delete. They sit alongside AI-generated icons on the same canvas, save with the rest of your project, and behave like any other layer.
Remove the background in one click
Select an uploaded image, look for "Remove background" in the right inspector, click. About ten seconds later, you have a transparent version of your image as a new version of the same layer.
A small but important difference
For uploaded images, the original is preserved as v1 and the background-removed version becomes v2 in the same family. You can switch between them in the version filmstrip. AI-generated icons replace in place; uploads keep history. We thought you'd want the original back if the bg-remove ever trims something it shouldn't.
It's the same one-click background removal you already use on AI-generated icons. Same quality, same one-credit cost, now applied to your own images.
Use any image as a starting point
This is the part we're most excited about. Select any uploaded image, type a prompt in the chat, hit Generate. Icon Maker uses your image as the visual base for the new generation.
There's no "use as reference" toggle. If you have an image selected and you generate, that image becomes the base. Drop a sketch of your icon idea, type "make this monochrome and minimal", done. Drop a photo of your product, type "turn this into a flat icon", same flow.
The same applies to upscaling and vectorization. Any utility you can run on an AI-generated icon, you can run on an uploaded image, and the upload is preserved as v1 throughout.
What's next
The drag-drop release is the foundation for two upcoming pieces we're already building:
- Provider-agnostic reference handling. Today, uploads pass through our FAL pipeline. We're working on a unified abstraction so style references route to Recraft, sketches to Gemini, and photos to FAL automatically, based on what you're trying to do.
- Brand kits. Drop your brand's images, lock in your palette, and have every generated icon match. Previewable later this quarter.
Try it now
Open the editor, drop an image, click Remove background. That's the whole loop.
Start uploading and let us know what breaks. Reply to any of our emails or ping us on Twitter. Your feedback shapes what ships next.
Thanks especially to the Icon Maker users who kept asking for this. This is the start of what you've been asking for.
FAQ
Stay in the loop
Get icon design tips, AI updates, and tutorials in your inbox.