Overview
HealAi is an AI-powered therapist platform designed to connect users with licensed therapists while providing always-available conversational support through a RAG-based AI assistant. The app helps people seek guidance, book sessions, and stay engaged between appointments — combining human care with intelligent, context-aware chat. Currently in development, HealAi is being built as a full-stack mobile product with realtime messaging, secure session management, and an AI layer grounded in trusted mental-health knowledge.
AI Chat & RAG Pipeline
Conversational Support, Human Care When Needed
The core idea of HealAi chat is simple: have a natural conversation with the user, understand what they need, and guide them — then connect them with a therapist when human support is the right next step.
We implemented Retrieval-Augmented Generation (RAG) so the assistant does not rely on free-form guesses. User messages trigger semantic search over a curated knowledge base: documents are chunked, turned into vector embeddings, stored in a vector store, and the most relevant passages are retrieved as context before the LLM generates a reply.
That means every answer is grounded in retrieved knowledge — more accurate, safer for mental-wellness topics, and ready to hand off to a live therapist through realtime Socket chat when the user needs deeper care.
| RAG Step | What It Does |
|---|---|
| Knowledge ingestion | Trusted wellness content is prepared and indexed for retrieval |
| Chunking | Documents are split into meaningful passages for precise lookup |
| Embeddings | Chunks become vectors for semantic similarity search |
| Retrieval | Top-matching context is pulled for the user’s query |
| Augmented generation | LLM replies using retrieved context + conversation history |
| Therapist handoff | When needed, the user is connected to a real therapist via chat |
Challenge & Solution
Problem
Accessing mental health support is often slow, fragmented, and intimidating. Users struggle to find the right therapist, wait too long for responses, and lack continuity between sessions. Generic chatbots without grounded knowledge can give unreliable answers, while traditional booking apps do little to support users in the moment they need help.
Solution
HealAi connects users with therapists through a modern mobile experience, while a RAG-powered AI chat assists with guided conversations using retrieved, relevant knowledge — not hallucinated responses. Realtime Socket-based messaging keeps therapist–user communication instant, and the NestJS + Prisma + AWS stack is designed for secure, scalable growth as the product moves from development toward launch.
Actively building core product flows and AI chat pipeline.
Connect users with therapists and manage sessions end-to-end.
Context-aware replies grounded in retrieved knowledge.
Live chat between users, therapists, and the AI assistant.
Connect With a Therapist
Video Call, Messaging & In-Person Sessions
Beyond AI chat, HealAi lets users connect directly with licensed therapists when they need real human support. From the therapist profile, users can start a conversation, join a live video session, or book a meeting — including in-person appointments.
- Messaging — continue care through secure 1-to-1 chat with your therapist
- Video call — meet face-to-face remotely for live therapy sessions
- Session booking — schedule upcoming appointments from the app
- In-person meetings — book offline sessions when meeting in person is preferred
Exercises for Relaxation
Guided Practices to Calm Mind & Body
HealAi includes relaxation exercises users can follow anytime — between sessions, after a tough day, or when they need a quick reset. These guided practices help reduce stress and support mental wellness alongside AI chat and therapist care.
- Breathing exercises — simple techniques to slow down and ease anxiety
- Mindfulness & grounding — short practices to stay present and centered
- Stretch & body release — light movement to release tension
- Daily calm routines — easy-to-follow sessions for consistent self-care
Realtime Chat (Sockets)
Socket-based realtime messaging powers ongoing conversations with the AI assistant and 1-to-1 chats with therapists — including typing indicators and instant message delivery without constant polling.
Key Features
- Connect users with therapists for guided mental wellness support
- RAG-based AI chat for context-aware, knowledge-grounded conversations
- Realtime messaging with Socket.IO
- Secure auth, profiles, and session-oriented workflows
- Mobile-first experience with React Native (iOS & Android)
- Scalable API layer with NestJS + Prisma
- Cloud deployment and services on AWS
My Role
- Architecting the full-stack platform (mobile, API, AI chat pipeline).
- Building the React Native mobile app for user and therapist experiences.
- Developing the NestJS backend with Prisma for typed data access.
- Integrating realtime chat using Socket.IO.
- Working with the Python RAG pipeline for embeddings, retrieval, and AI responses.
- Planning AWS deployment for APIs, media, and scalable services.
Tech Stack
Technical Implementation
Mobile: React Native (iOS & Android) | Backend: NestJS + Prisma | AI Chat: Python RAG pipeline (chunking, embeddings, retrieval + LLM) | Realtime: Socket.IO for live chat | Cloud: AWS for hosting, storage, and scalable services
| Layer | Technology |
|---|---|
| Mobile App | React Native |
| API / Backend | NestJS |
| ORM / Database | Prisma (+ PostgreSQL) |
| AI / RAG | Python RAG pipeline |
| Realtime Chat | Socket.IO |
| Cloud | AWS |
Deliverables (In Progress)
- React Native mobile app (user & therapist flows)
- NestJS backend API with Prisma
- RAG-based AI chat service (Python)
- Realtime Socket chat for user–therapist messaging
- Therapist discovery and connection workflows
- AWS infrastructure setup for development and staging
Project: HealAi | Industry: HealthTech / Mental Wellness | Status: In Development
