Start labeling
Open a local dataset or load the bundled sample workspace.
Folder access is unavailable here; use the sample workspace.
Open a local dataset or load the bundled sample workspace.
Ctrl + wheel: zoom
Ctrl / Space + drag or Middle drag: pan
Open a dataset to outline samples.
Name a target, then outline it on the reference image.
Not saved
npm run yoloe:setupnpm run yoloe:preparenpm run yoloe:startlabel folder. Create it now to save annotations in this dataset.
| ID | Class Name | Color | Actions |
|---|
Change Detection box IDs while keeping box positions and sizes. Class file names, colors and saved Layout presets stay unchanged.
Use +3 to increase or -3 to decrease. Every resulting ID must be 0 or greater.
| Original class | New class | Boxes | Actions |
|---|
Count shows boxes in the current image, including hidden classes. Each mapping uses the original ID: 0 → 2 and 2 → 0 swap classes. Unlisted IDs stay unchanged. Multiple original IDs may share a new ID.
| Original class | New class | Boxes |
|---|
Enter a class ID manually or select from the available classes below.
Easy Labeling is designed for desktop browsers like Chrome or Edge that support the File System Access API.
For the best experience and full functionality, please access this page from a desktop or laptop computer.
Mobile devices and some browsers are not currently supported.