Inspiration
The increasing prevalence of social isolation worldwide, even more so after the pandemic, inspired me to create MIA, an AI companion designed to provide emotional support through natural conversations.
What it does
MIA offers judgment-free conversations using voice interaction, learning from each interaction to become a personalized emotional support companion. Features include:
- Voice recognition (STT, ASR) and synthesis (TTS)
- Emotional response system
- Short (current conversation) and Long (past interactions) term memories
- Visual feedback
How we built it
Backend: n8n for workflow automation Database: Supabase to store message history Frontend: Plain HTML, CSS, and JavaScript
Challenges we ran into
Limitations of the Browser Speech API Managing real-time response latency Synchronizing voice input, visualization, and webhook requests
Accomplishments that we're proud of
Seamless voice interaction and cross-browser compatibility Cool Visuals Memory management pipeline
What we learned
Building LLM-powered backends with no/low-code tools Frontend-backend integration Database management
What's next for Mindful Interactive Assistant
Fully local implementation User authentication Encrypted communication
Built With
- css3
- html5
- javascript
- n8n
- supabase
- webhook
- webspeechapi
Log in or sign up for Devpost to join the conversation.