Care Archive

Building the data infrastructure behind care robotics.

Care Archive is an African data lab modelling human embodied intelligence in care. We ethically capture how humans actually care for other humans in South African homes and other spaces - across diverse skin tones, body types, languages, and physical and emotional environments.We pair this with the clinical annotation, consent infrastructure and privacy frameworks that robotics labs teams around the world require.The training data behind real-world care robots does not yet exist at scale. Care Archive helps robotics labs meet the global eldercare crisis with the robotic care assistants the world urgently needs.

On care, culture and what we capture

African and other Global South care traditions - multi-generational, embodied, respectful of elders - are already shaping how care is delivered in high-income countries. Many frontline care workers in London care homes, New York hospitals, and Sydney private residences come from African, Filipino, and South Asian backgrounds. The cultural transfer in care is not theoretical. It happens every day.Care Archive treats this as a strength.We don't try to capture care as a cultural practice. Care is relational, emotional, irreducibly human. We capture the physical and biomechanical layer underneath it: how a body bears weight when lifted from a chair, the angle of grip when dressing a wound, the navigation of cluttered domestic space, the recognition of distress in posture. These are physical problems that current AI fails at because its training data is overwhelmingly Western, institutional, and demographically narrow.The cultural and relational dimensions of care remain - and should remain - irreducibly human. The physical reality of caring for diverse types of ageing bodies in diverse types of homes is what we work to make legible to the AI systems that will increasingly shape global care.

Annotated, synchronised, training-ready

Phase 1 captures:

Phase 2 expands to tactile force maps and IMU-based motion capture, validated with robotics research partners.Anonymised by default. Every participant signs explicit, commercial-use consent. Compliant with POPIA, GDPR, and HIPAA frameworks.

Reversing the flow of AI innovation

The AI industry has a documented history of taking from the Global South - labour, data, attention - without giving back. Care Archive is built deliberately against that pattern.

We design and market only that which has been freely given, fairly compensated, and properly understood.

Get in touch

For robotics labs, researchers, partners, and press:

• [email protected]

Cape Town, South Africa