Healer-AI

Inspiration

We've all seen it—a family member struggling to remember their symptoms at a doctor's appointment, or a doctor looking buried in paperwork instead of engaging with their patient. The healthcare journey, especially in a place like India, is full of these small frictions that add up. We were inspired to build something that could ease this burden for both patients and doctors, using AI to handle the tedious parts so humans can focus on healing and being heard.

What it does

HealerAI is an AI assistant that streamlines the entire clinical process.

Pre-Consultation: It starts with an AI-driven intake call that schedules an appointment and creates a preliminary summary for the doctor.
Consultation: During the visit, it can process the conversation to generate structured clinical notes automatically.
Post-Consultation: It powers a smart chatbot, allowing doctors to quickly recall case details and patients to get reliable answers to follow-up questions, cutting through the noise of online misinformation.

How we built it

We orchestrated a multi-agent AI system using Langflow as the central nervous system. The workflow uses different large language models for specific tasks: Gemini for orchestrating the flow and grok for generating realistic medical conversations and reports. We used Supabase for our database and serverless functions to create a robust API, and AstraDB as the vector database to give our AI a long-term memory for patient conversations.

Challenges we ran into

Getting the AI to sound genuinely human and medically sound was a huge challenge. We spent countless hours refining prompts to ensure the conversations were natural and the medical summaries were accurate and professional. Integrating the different services—Langflow, Supabase, AstraDB, and multiple LLM APIs—into one seamless, deployable system also took a lot of effort to get right.

Accomplishments that we're proud of

We're incredibly proud that we built a complete, end-to-end system, not just a proof-of-concept. Creating a multi-agent workflow where each AI has a specific job—from receptionist to clinical scribe—and having it all work together is a major accomplishment. The post-consultation RAG agent, which allows doctors and patients to "query" past conversations, feels like a genuinely useful tool that could make a real difference.

What we learned

This project was a deep dive into the power of agentic AI. We learned that by breaking down a complex problem into smaller tasks and assigning each to a specialized AI agent, you can build incredibly powerful and nuanced applications. We also learned firsthand that the quality of an AI system is directly tied to the quality of its prompts; careful prompt engineering is everything.

What's next for Healer-AI

The current system is a powerful backend, but the next logical step is to build a user-friendly interface for it—a web or mobile app for doctors and patients. We also plan to integrate real-time voice transcription to make the consultation logging seamless and explore using voice agents like Eleven Labs to fully automate the initial patient intake calls.