Case study · AI & automation
A virtual nurse you can talk to
Nurse AI Hub helps nurses and healthcare learners build knowledge and confidence. We built a virtual nurse that holds a spoken conversation, answering medical questions or acting as a partner for practising medical English.
- Client
- Nurse AI Hub
- Sector
- Healthcare education
- Engagement
- AI product build
- Delivered
- 2024
- In and out
- Voice
- Users can speak their question and hear the answer read back with text-to-speech.
- Avatar responses
- Live
- An on-screen avatar responds in real time, so the exchange feels like a conversation rather than a form.
- Use cases
- 2
- Asking health-related questions, and practising medical English in a realistic setting.
What it had to do.
Nurses and students preparing to work in English-speaking healthcare need to practise the conversations they will actually have. Reading a textbook does not help with speaking, and a typed chatbot does not either.
The product had to feel like talking to a colleague: speak a question, hear an answer, and see someone respond, while staying focused on medical topics.
The system we built.
A Django service handles the conversation. A React client turns it into something you speak to and watch respond.
- ListenVoice input captures the user's question, with text input available as well.
- AnswerThe assistant answers in the role of a virtual nurse, scoped to medical questions and medical English practice.
- SpeakResponses are read aloud with text-to-speech.
- Respond on screenAn avatar reacts in real time as the answer plays, completing the conversational loop.
Live at nurseaihub
The public site introduces the learning modules. The assistant itself sits behind a free account.
- Backend
- Python · Django · DRF
- Frontend
- React
- Interaction
- Voice · Text-to-speech · Avatar
- Access
- Sign-in required
- Status
- Live
Want an assistant people talk to, not type at?
Bring the conversations your users need to have. We will tell you which parts need voice, which need guardrails, and what to build first.
30 minutes · No pitch · A written summary afterwards