Elbrain

AI-powered mental health companion offering smart, adaptive assessments for depression, anxiety, ADHD, PTSD, and more — all in a simple, fun, and stigma-free experience.

Comment

Inspiration

Mental health struggles are universal, yet access to effective, stigma-free assessments is still limited. We realized that most current tools feel cold, impersonal, and overly clinical — creating a barrier for those seeking help. We wanted to build something better: an experience that feels conversational, safe, and human — powered by AI.

That’s how Elbrain was born: an intuitive mental health companion designed to assess conditions like depression, anxiety, PTSD, ADHD, and eating disorders using adaptive questioning, all while ensuring the user feels understood and supported.

What it does

Elbrain is an AI-powered mental health assessment tool that:

  • Asks dynamic, intelligent questions based on user responses
  • Evaluates for multiple mental health conditions in a single flow
  • Generates a personalized report with insights and recommendations
  • Uses a friendly, engaging tone to reduce the feeling of taking a clinical test
  • Prioritizes user privacy and non-judgmental interaction

The goal is to provide early detection and guidance while making mental health check-ins feel natural and stigma-free.

How we built it

We started by mapping mental health diagnostic criteria based on clinical standards like the DSM-5. Then, we designed an adaptive questioning engine using:

  • LLMs (like LLaMA 3 and GPT-4) for generating intelligent, empathetic questions
  • A scoring logic that updates condition likelihoods in real time
  • A frontend built in React.js with a calming, conversational UI
  • FastAPI for backend APIs and data routing
  • MongoDB for storing user data and assessment states securely
  • LangChain to orchestrate the LLM prompts
  • Deployment via Vercel (frontend) and Render (backend)

Challenges we ran into

  • Designing prompts that feel empathetic yet structured
  • Preventing the AI from making diagnostic claims while still offering useful insights
  • Building a balanced scoring mechanism that reflects mental health nuances
  • Ensuring user responses remain confidential and secure
  • Making the interface feel engaging but not overwhelming

Accomplishments that we're proud of

  • Creating a multi-condition assessment that feels personal and intuitive
  • Implementing adaptive logic to tailor questions based on user mood and tone
  • Building a mental health tool that doesn’t feel like a test — but a conversation
  • Successfully combining generative AI with clinical frameworks
  • Making mental health tools feel more accessible and stigma-free

What we learned

  • Users are more honest and open when language feels friendly
  • Context-aware questioning significantly improves assessment quality
  • Generative AI can help simulate real empathy — when guided correctly
  • Building for mental health requires a deep respect for boundaries, tone, and ethics
  • Even small UI/UX changes can drastically improve user comfort and trust

What's next for Elbrain

  • 🧠 Gamifying the experience to further reduce test anxiety
  • 🌍 Localization and multilingual support for global reach
  • 🏥 Partnering with mental health professionals and institutions
  • 📈 Adding analytics to track well-being over time
  • 🔐 Implementing end-to-end encryption and full anonymization
  • 🤝 Integrating with corporate wellness platforms for team mental health solutions

We're just getting started — our goal is to make mental health tools more human, proactive, and accessible for everyone.

Built With

  • fastapi
  • langchain
  • llama3
  • mongodb
  • vercel
Share this project:

Updates