-
-
Analysis
-
Tips
-
Dashboard
-
History
-
About
-
Privacy
🎭 VibeMaster: Bridging AI and Emotional Intelligence
Inspiration
In an era of digital burnout, understanding our emotional "vibe" is crucial for productivity. We were inspired to build VibeMaster to prove that AI can be more than a cold logic engine—it can be a tool for self-reflection and mental wellness. We wanted to create a bridge between complex NLP (Natural Language Processing) and an accessible, beautiful user experience.
How i built it
The project was developed following the Vibe Coding philosophy, focusing on rapid iteration and high-quality UI/UX.
- Frontend: Built with Streamlit, enhanced with custom CSS for a modern "Glassmorphism" aesthetic.
- AI Engine: We integrated the Groq API to leverage the Llama 3.3 70B model, chosen for its ultra-low latency and sophisticated reasoning.
- Data Flow: The app processes raw text into structured JSON data, allowing us to visualize mood trends and generate dynamic recommendations in real-time.
Challenges we faced
One of the main hurdles was ensuring the LLM consistently returned valid JSON for our dashboard's metrics. We overcame this through advanced prompt engineering and regex-based fallback mechanisms. Another challenge was balancing the visual complexity of the custom CSS with Streamlit's native components to ensure a responsive "Vibe".
What we learned
We deepened our knowledge in:
- API Integration: Mastering high-performance inference with Groq.
- Sentiment Analysis: Understanding how to calibrate polarity and emotional weight in text.
- Prompt Engineering: Learning how to force LLMs to act as specialized wellness coaches.
Future Plans
We aim to integrate ElevenLabs for voice-based affirmations and connect with Spotify's API to generate real-time playlists based on the user's "Vibe Score" history.
Log in or sign up for Devpost to join the conversation.