LLMs reached the quality threshold
Until 2023, conversational AI lacked the nuance required for sustained health coaching. Large language models now enable context-aware, emotionally adaptive conversations that improve with every session.
Aura is an AI coaching companion for people with Type 2 diabetes and their families, learning from every conversation to provide continuous, personalized support between clinical appointments.
Every digital health tool built for diabetes focuses on the individual patient. But diabetes self-management happens at the dinner table, in shared grocery runs, and in the small daily decisions made inside a household.
Family members want to help. Without guidance, that help often becomes control, creating the friction researchers call "diabetes nagging". Our user study found this dynamic to be one of the most consistent sources of stress for patients managing T2D with family present.
No existing solution addresses both sides of this relationship with separate, role-appropriate experiences. Aura does.
Users valued autonomy-preserving support, transparency about what the AI knows, and the reduction of interpersonal conflict around diabetes management.Master's Thesis, Uppsala University, 2026
Patient controls the bridge
structurally, not just by policy.
Until 2023, conversational AI lacked the nuance required for sustained health coaching. Large language models now enable context-aware, emotionally adaptive conversations that improve with every session.
Remote and digital care became normalized globally. The expectation that health support exists between appointments is now established; the infrastructure to deliver it is not.
GDPR and growing data rights awareness mean that privacy-by-design is no longer optional. Aura's structural privacy boundary, built into the architecture rather than bolted on, becomes a competitive advantage as regulation tightens.
Design Science Research, Master's Thesis, 2026
The dual-role architecture, privacy mediation mechanism, and coaching philosophy behind Aura were developed through a Design Science Research methodology at Uppsala University's Department of Information Technology.
The prototype was evaluated through scenario-based walkthroughs and think-aloud protocols with patient-supporter dyads, using semi-structured interviews and the System Usability Scale (SUS).
Progress matters more than hype. This is the actual state of the company and product today.
Uppsala University Master's Thesis published April 2026. Qualitative user study with patient-supporter dyads complete.
Full-stack prototype: React Native app, Express/TypeScript backend, Supabase, AI agent with 4 specialized flows. Healthcare-grade architecture: PHI-safe logging, Zod-validated tool envelopes, Row-Level Security.
Seeking first clinical pilot partners in Sweden. Waitlist open for early users.
Exploring reimbursement routes via Swedish Tandvårds- och läkemedelsförmånsverket (TLV) and EU MDR Class I pathway for wellness applications.
Sweden has 500,000 people living with diabetes and one of the world's most digitally mature healthcare systems. The National Diabetes Registry (NDR) demonstrates the system's openness to data-driven care. Region partnerships and vårdcentral distribution channels are a clear path to scale without direct-to-consumer acquisition costs.
The family-aware architecture is difficult to replicate. The privacy boundary is structural, a technical design decision rather than a policy. The dual-role coaching system requires purpose-built AI agent flows, a typed memory architecture, and behavioral science grounding that cannot be added to a general chatbot.
We are raising a pre-seed round to fund clinical pilot partnerships, regulatory pathway assessment, and the team needed to go from prototype to product. If you invest in digital health or AI infrastructure, we would welcome a conversation.
Aura is in active development. We are onboarding our first users in Sweden in 2026. Join the waitlist to be among the first invited.