
The easiest & fastest way to annotate images for vision AI. | Created by @wkentaro_ | 16K GitHub stars
Joined November 2023
- Tweets393
- Following2
- Followers340
- Likes1.3K
Turn it on in File › Connect AI Agent… and click Add to Claude Desktop. ChatGPT and Claude Code take one click too.
Your AI agent can now annotate in Labelme.
It looks at the image you open, reads existing shapes, adds new or deletes wrong ones. You approve each change.
Fit to Selection zooms to each shape you select. Turn it on with Ctrl+Shift+F and step through the Shape List with the arrow keys.
Labelme v7.9 is built for reviewing annotations.
Lock each shape you have checked with Ctrl+Shift+L. Labelme won't move, reshape, relabel or delete it until you unlock it. The lock is saved in the JSON file.
You can now download and manage AI models right in Labelme’s Settings.
Keep annotating while models download in the background. Included in Labelme 7.6.
Launch Labelme with nothing open and it now shows where to start: Open Image, Open Directory, or drop files onto the canvas
New in Labelme v7.5:
- Merge masks that share a label. One Undo brings them back
- An empty window shows where to start
- Huge images open after a confirmation
Plus fixes for Undo, Label List reordering, and zoom
Use one color for every shape, or assign colors automatically by label
Three polygons, One plant
Group ID keeps disconnected parts together





