Meet a quicker, steadier
Brimsund research model
Try Brimsund
Read the notes
Latest from the lab
All releases
Segmentation
SKARVA 3
Traces moving subjects across 41 minutes of footage without a single re-prompt.
Vision backbone
NAUSTv3
Self-supervised features that hold up on grainy, low-light survey frames.
World model
V-EBBA 2
Predicts the next 3.2 seconds of a scene well enough to plan against it.
Brimsund
Platform
Model Lab
Developers
Company
Try Brimsund
Model Lab
Overview
Focus Areas
The Team
Releases
SKARVA 3
NAUSTv3
V-EBBA 2
All releases
For Builders
Toolkits & SDKs
Papers
Open Source
NAUSTv3
tracking
NAUSTv3