labelImg Next

Browser-based image and video annotation. Supports YOLO and Pascal VOC. App: 127.0.0.1:8000.

Local installation

Requires Python 3.12+ and Node.js 22.

Windows

Double-click start.bat, or run:

.\start.bat

setup.bat prepares dependencies without launching; install_sam.bat installs or repairs SAM.

Ubuntu

sudo apt install python3-venv python3-tk
python3 -m venv .venv
.venv/bin/python -m pip install -r requirements.txt
bash ./build.sh
.venv/bin/python main.py

Docker

Copy .env.example to .env and set LABELIMG_DATASETS_DIR to your dataset folder.

# Published image
docker compose -f docker-compose.ghcr.yml up
# Or build from source
docker compose up --build

Open /datasets through File → Backend path.

Video

Open a dataset, then Video. Save annotated frames to the dataset; empty frames require Save as background.

Videos stay in their original folders; progress is stored in data/video-metadata/. Reselect a browser-selected video to resume. Deleting it does not remove saved frames.

SAM

Requires an NVIDIA GPU, a CUDA 12.6+ driver and Hugging Face model access. Use Plugins to install the runtime and download weights. Windows launchers prepare the runtime automatically.

Development

With .venv activated (npm.cmd instead of npm in Windows PowerShell):

python main.py --reload
npm run dev --prefix frontend   # Separate terminal; http://127.0.0.1:5173
python -m pip install -r requirements-dev.txt
python -m unittest discover -s tests -v
npm test --prefix frontend