Label the World in 3D

Tomorrow's robots, computer vision and augmented reality applications require precise algorithms and lots of quality data. Stray Studio allows you to annotate entire scenes in one go, instead of having to label data one image frame at the time.

Stray Label Studio for annotation spatial AI data.
Stray Scanner app scanning a scene.

Bring Your Data

Import your data from a depth sensor or stereo camera, or use our Scanner app.

Stray Studio will 3D reconstruct the scene from your dataset.

Annotate in 3D

Use the Stray Studio desktop app to annotate your scene. Paint objects for semantic segmentation. Label points of interest with keypoints. Add bounding boxes around objects. Associate keypoints with vectors or coordinate frames. Define object poses.

Generate your datasets

From the annotated scene, Stray Studio can generate labeled examples for each frame in your dataset. Semantic segmentation maps. Completed depth maps. Heatmaps for keypoints. Feature maps for vectors. Bounding boxes. 6D poses.

Set up training and pipelines

Use the labeled data to train your model to predict the labels. You can easily import the for model training using the Stray CLI.


Stray Studio is part of the Stray toolkit, which can be installed from here.

Subscribe to Updates

If any of this sounds exciting and you would like to stay tuned for more, subscribe to our newsletter.

We will only email you with significant updates to our tools and might sometimes include related links we find useful from across the internet. You can unsubscribe at any time.

Subscribe Here