The AI Hackers

Inspiration

We have always found journaling helpful, but sometimes we don’t know what to write, or we're unsure how to express our feelings. We wanted to create a simple tool that could gently guide people toward self-awareness and reflection. With the rise of AI, we thought: What if a virtual friend could help people understand their emotions just by reading what they type?

What it does

How we built it

Frontend (React): A simple UI where users can write about their day. Displays detected mood (e.g., happy, stressed, sad) after journal submission. Shows suggestions (like a quote or calming activity) based on mood.

Backend (Flask or Node.js): Sends the journal entry to a sentiment analysis API. Receives mood label (positive, neutral, negative). Uses basic rules to recommend a quote, playlist, or activity.

AI/ML: Used a pre-trained sentiment model (like textblob, VADER, or OpenAI). No need to train anything—just plug and play!

Challenges we ran into

Sometimes a message could sound positive but be sarcastic. I had to test several models to get decent results and manually tweak some responses. We didn’t want to make a full-blown mental health app, so we had to keep the language neutral and avoid giving advice that sounded too serious or medical. Making the app feel warm, friendly, and non-intrusive was harder than expected. We learned a lot about user experience design during this process.

Accomplishments that we're proud of

We were able to utilize AI and ML enable users to: Write a quick journal entry each day. See how their mood is being interpreted. Get an AI-generated suggestion tailored to their mood.

What we learned

During the project, we learned: How to use OpenAI's API (or Hugging Face models) to analyze sentiment in text. How to build a basic React frontend and connect it to a Flask backend. How to handle user input safely and provide AI-generated responses. How to keep things lightweight and privacy-conscious (nothing is saved unless the user wants to).

What's next for The AI Hackers

We plan to: Add more personalized suggestions (e.g., recommend a nearby park if the user is feeling overwhelmed). Allow users to see mood trends over time via charts. Add voice journaling with speech-to-text support.

Built With

• css
• javascript
• python
• react
• render