Inspiration
Mental health and emotional well-being are increasingly important in our fast-paced world today. Many people struggle to find immediate, empathetic support when they need it most. Traditional therapy can be expensive, hard to access, or have long wait times. I was inspired to create Eli - an AI-powered emotional support companion that provides immediate, personalized, and compassionate assistance 24/7.
What it does
Eli is an emotional support application that uses two AI agents working in collaboration to provide comprehensive mental health support. Users can chat naturally about their feelings, struggles, or concerns and Eli will provide compassionate responses based on analysis of the conversation context and emotional state of the user. It also suggests self-help tips and techniques.
How I built it
I built Eli using a multi-agent architecture with Google's Agent Development Kit (ADK) approach. The front end is in React and backend in Python and Google Gemini has been used as the backbone LLM of choice.
Challenges I ran into
Some Cloud Build Issues along with Billing and Permission issues I encountered while deploying to Google Cloud.
Accomplishments that I am proud of
I am proud of Eli and it's immense potential in an increasingly complex fast-paced world.
What I learned
I learned how to design and implement specialized AI agents that work together, sharing context and collaborating on complex tasks. Discovered patterns for coordinating multiple AI agents, including context passing between agents, structured data formats for agent communication, workflow design for multi-agent systems.
What's next for Eli - Emotional Support
Persistent conversation memory across sessions, more nuanced emotional state detection, automatic detection of crisis situations, additional specialized agents for specific scenarios, learning user preferences and communication style
Log in or sign up for Devpost to join the conversation.