The Problem
Mental health is a major issue in the U.S., especially among teens—15% experience symptoms of depression. Many face barriers to accessing support, including stigma and cost. A chatbot offers a scalable solution, providing accessible, real-time mental health assistance.
What it does
Our chatbot enables real-time conversations with a D-ID-powered virtual therapist. It leverages face-api for emotion detection, integrating facial cues into the dialogue to enhance the therapeutic experience. The system uses a fine-tuned model and an embeddings framework to match users with the appropriate therapy style, supported by a retrieval-augmented vector database of therapy transcripts. The result is personalized, human-like therapy sessions that teens can access in a supportive digital space.
How we built it
We combined face-api for facial emotion detection, Google microphone input for voice interaction, ChatGPT-4 for AI-driven responses, D-ID for the therapist avatar, and ElevenLabs for voice generation. Our vector database, managed with SQLite, houses therapy transcripts from public sources. The frontend is built with HTML, CSS, and React, creating an intuitive interface for users.
Challenges we ran into
Integrating D-ID's avatar with ChatGPT’s responses was our biggest hurdle, requiring careful coordination to ensure smooth, natural conversations.
Accomplishments that we're proud of
We’re proud to have created a fully functional solution that combines multiple complex technologies. While there's room to grow, we've taken a meaningful step toward solving the mental health challenges faced by teens today.
Log in or sign up for Devpost to join the conversation.