"Yasasii Call" to Reduce Mental Strain

Project: Yasashii Call

Self-reported purpose: An AI-powered nurse call system designed to reduce mental strain on nurses and patients by translating patient requests into actionable summaries while preserving the original message.

Key claim: The system supports patients who hesitate to ask for help, reduces the burden on nurses interpreting incomplete or emotional requests, and does not replace human judgment or diagnosis.

What changed: The project evolved from a simple nurse-call translator into a broader hospital task-routing system, but was pared down to a working prototype demonstrating core functionality.

Most important open question: Does the described prototype demonstrate sufficient commercial viability or traction to warrant further due-diligence attention?

The description is self-reported and unverified. No evidence of revenue, customers, funding, or adoption is provided. The project appears to be an early-stage prototype built for a hackathon.

• The description states that Yasashii Call is an AI nurse call system.
• It allows patients to describe what they need in their own words.
• GPT-5.6 Luna translates the message into a short, actionable summary for the nurse dashboard while preserving the patient’s original words.
• Nurses can accept, forward, complete, reopen, and review calls through a simple workflow.
• The AI does not diagnose, assign medical urgency, or recommend treatment.
• The system includes a patient interface, nurse dashboard, call-status workflow, local history, responsive design, API route, fallback behavior, and deployment.
• It was built using Codex and the OpenAI Responses API with GPT-5.6 Luna.
• The prototype is deployed as a public web demo.

Inference: The system appears to be a communication tool that enhances nurse-patient interaction by structuring patient requests for easier processing, without replacing clinical decision-making.

• The description states the project was inspired by the observation that patients hesitate to call nurses due to uncertainty about ward busyness.
• It also notes that nurses must interpret incomplete or emotional requests while managing competing tasks.
• The system is positioned as a way to reduce mental burden on both sides without replacing human communication or medical judgment.

Inference: The positioning evolved from an idea of helping with nurse calls into a more comprehensive hospital task-routing system, but was simplified for demonstration in a hackathon setting.

• The description states that the system targets nurses and patients in hospitals.
• It aims to support patients who hesitate to ask for help.
• Nurses are described as needing to interpret incomplete or emotional requests while managing many competing tasks.

Not evidenced: No specific customer segments, personas, or use cases beyond general hospital settings are provided. No evidence of target market size, segmentation, or prioritization.

• The description does not state any pricing model or business model.
• It mentions that the prototype was deployed as a public web demo and its source code and documentation were published on GitHub.
• There is no indication of monetization, licensing, or customer acquisition strategies.

Not evidenced: No evidence of revenue streams, pricing, or commercialization plans beyond a hackathon prototype.

• The system was built using Codex to turn the initial concept into a working vertical prototype.
• It uses the OpenAI Responses API with GPT-5.6 Luna for message translation.
• Features include patient interface, nurse dashboard, call-status workflow, local history, responsive design, API route, fallback behavior, and deployment.
• The project was deployed as a public web demo.
• Source code and documentation were published on GitHub.

Inference: The prototype demonstrates basic software engineering capabilities and integration with AI APIs. It is not clear if the system has been tested in real-world environments or scaled beyond a demo.

• The project was submitted to the OpenAI 2026 hackathon.
• It includes a public web demo and GitHub repository.
• The authors note that this was their first time building and publishing a web application.
• No evidence of user adoption, customer feedback, or post-hackathon development is provided.

Not evidenced: No data on users, usage metrics, or product maturity beyond the prototype stage.

• The description does not mention any existing competitors or similar systems.
• It does not reference prior art in nurse call systems or AI-assisted hospital workflows.
• No market analysis or competitive positioning is provided.

Not evidenced: No evidence of competitive landscape, market gaps, or differentiation from existing solutions.

• The system is described as a prototype built for a hackathon with no clinical validation or safety review.
• It explicitly states that the AI does not diagnose or recommend treatment, but it remains unclear how this is enforced in practice.
• The project was built by one person (team size: 1), which raises questions about scalability and long-term development capacity.
• No evidence of regulatory compliance, privacy safeguards, or integration with real hospital systems is provided.

Inference: The lack of clinical testing, safety protocols, and commercial viability raises significant concerns for any future investment or partnership.

1. What specific feedback have you received from healthcare professionals or end-users?
2. How do you plan to ensure the AI output remains a communication aid rather than a decision-making tool in real-world use?
3. Have you considered regulatory compliance, data privacy, and security requirements for healthcare systems?
4. Is there any intention to collaborate with hospitals or medical institutions beyond the prototype stage?
5. What are the technical and operational challenges that would need to be overcome before deploying this system at scale?