Access to primary care is important. But currently, it sucks. While trying to identify a core problem to solve, we discovered that a significant percentage of the users we spoke to were frequent healthcare users, indicating a high need for faster access to care. Sixty-two % of them said they were looking for credible tools to discuss health symptoms in the context of their medical history. Patients often feel that getting in touch with a Primary Care Physician is time-consuming, inconvenient, and awkward.
What it does
gendoc.ai delivers instant, personalized responses and centralizes healthcare in one place. Strengthening the patient–doctor loop and making it easier to reach out for potentially life-saving help.
We used Figma to prototype the design spec and visualize the key components of our MVP, which would be directly interacting with our end client
We built a FastAPI server acting as an intermediate server between the client and agent. The FastAPI will handle transactional processing for our application and communicate with the MCP Agent server to perform tasks and respond to users promptly
Claude MCP Agent: We built a main agent that will handle communication and orchestration. Depending on our system prompt + user prompt, our agent offers a tool to:
Communicate with the patient to understand their needs
Predict symptoms based on the user's input
Generate a report and get the doctor's approval
Place the medicine order according to the doctor's request
Schedule an appointment with a specialized doctor if necessary
Accomplishments that we're proud of
Ease of use due to the user-friendly interface
Quality of the insights generated by our model
What we learned
How to build an MCP agent with the Claude Agent SDK
Log in or sign up for Devpost to join the conversation.