// the find
HumanSignal/labelImg
LabelImg is now part of the Label Studio community. The popular image annotation tool created by Tzutalin is no longer actively being developed, but you can check out Label Studio, the open source data labeling tool for images, text, hypertext, audio, video and time-series data.
LabelImg is a desktop GUI tool for drawing bounding box annotations on images, saving them as PASCAL VOC XML, YOLO txt, or CreateML JSON. It was the go-to choice for small object detection datasets for nearly a decade. It is now officially abandoned — the repo was absorbed by HumanSignal to funnel users toward Label Studio.
The keyboard shortcut workflow (w/a/d for draw/prev/next) is genuinely fast for solo annotators doing bounding boxes. YOLO format output works without any conversion step, which matters when you just want to feed images straight into a training script. It is a single Python file with Qt — no server, no database, no account, runs completely offline. The codebase is small enough that you can read and understand all of it in an afternoon.
It is dead software. Last real commit was years before the June 2024 push, and that push just updated the README to redirect you elsewhere. Only bounding boxes — no polygons, no segmentation masks, no keypoints — so you will hit a wall the moment your task needs anything beyond rectangles. The installation story on macOS and Windows involves multiple manual steps with pyrcc5 and Qt version pinning that break constantly as OS and Python versions advance. Label Studio, the suggested replacement, is a full web application that requires running a server, which is a significant operational jump for someone who just wanted a desktop tool.