# ST-VLA

**ST-VLA** is a vision-language action prediction model augmented with spatial traces for robotic manipulation.

More details coming soon.