Matcha Journal

An AI-powered journaling companion that listens with empathy, tracks your mood, and helps you reflect through thoughtful conversation.

Comment

Inspiration

I wanted to create a safe, beautiful space where anyone could journal freely without judgment. Mental health and self-reflection matter deeply, and I felt AI could make that experience warmer and more personal than writing alone.

What it does

Matcha Journal is an AI-powered journaling companion. You log in with your name, select your current mood, and start writing. The AI responds with empathy and asks gentle follow-up questions to help you reflect deeper. Every entry is saved to a database and your mood history is tracked over time with visual analytics.

How we built it

Built entirely in Python over one weekend using Streamlit for the frontend with custom CSS for the matcha green aesthetic, animated walking teddy bears, and floating leaves. Groq API powers the LLaMA 3.3 70B model for fast empathetic AI responses. MongoDB Atlas stores all journal entries and mood data. Deployed live on Streamlit Cloud.

Challenges we ran into

Managing multiple Python versions on Windows caused several dependency conflicts. Getting the Gemini API to work failed due to quota issues so we switched to Groq mid-build. Keeping API keys secure while deploying required setting up Streamlit secrets properly. Getting the CSS animations to work smoothly inside Streamlit took significant trial and error.

Accomplishments that we're proud of

Built and deployed a fully working AI app in under 24 hours. The matcha aesthetic with walking teddy bears turned out genuinely beautiful. The AI responses feel warm and human rather than robotic.

What we learned

How to build and deploy a full stack AI app in a weekend. How to use the Groq API for fast LLM inference. How to design a calming aesthetic UI using only Python and CSS. How to handle secrets securely in cloud deployments.

What's next for Matcha Journal

Adding user authentication with passwords so entries are truly private. Sentiment analysis on entries over time. Weekly mood reports sent by email. A mobile friendly version.

Try it out

Live App | GitHub Repository

Built With

Share this project:

Updates