Get accurate image annotations with Labelme, a Python tool for founders, with 16k+ GitHub stars.
16,035 stars3,689 forksPythonQuality 8/10Updated 7/10/2026100% free ยท open source
What it does
Labelme is a Python tool for accurately annotating images with various shapes and AI-assisted features to support machine learning model training.
Install / run
pip install labelme
When to use it
โขWhen training a machine learning model that requires annotated images
โขWhen creating a dataset for object detection or image classification tasks
โขWhen needing to annotate images with specific shapes like polygons, rectangles, or circles
Quick start
1Run the tool using the command `labelme` to open the annotation interface
2Load an image by selecting 'Open' and choosing the desired image file
3Create a new annotation by selecting a shape tool (e.g., polygon, rectangle) and drawing on the image
4Save the annotated image by selecting 'File' > 'Save' and choosing the output format (e.g., JSON, PNG)
5Use the `--labels` option to specify a list of labels for annotation, e.g., `labelme --labels labels.txt`
Ready-to-paste prompt
labelme --labels labels.txt image1.jpg
Heads up: The AI-assisted annotation feature requires a pre-trained model, which can be downloaded using the `labelme --download_model` command, but this may require additional dependencies or setup
Saves to your device
Topics
annotations
classification
computer-vision
deep-learning
image-annotation
instance-segmentation
python
semantic-segmentation
video-annotation
What's inside โ free to inspect
No purchase needed
Read the entire source before you build โ unlike paid marketplaces that hide it behind a buy button.