Animus Labs
Teaching machines how humans work.
Physical AI has a data problem. We're solving it at the source.
About
Animus Labs builds the training-data layer for embodied AI. We capture first-person, multimodal recordings of real people doing real work in live industrial environments across the EU, then turn that footage into rights-cleared, annotated datasets that humanoid and robotics teams can train on directly.
Every hour we deliver comes with full provenance: documented consent, EU jurisdiction, and a compliance record that travels with the data.
We're currently in private pilots with a small number of partners.
Real environments, not labs.
Data from the floors where robots will actually be deployed.
Multimodal by default.
Vision, depth and motion captured together, ready for modern VLA pipelines.
Provenance built in.
GDPR-native from capture to delivery, with consent and erasure handled at the source.