Inspiration
Millions of people, especially young adults, silently struggle with gambling addiction—often enabled by online betting apps and fantasy sports. This leads to financial ruin, emotional distress, and broken relationships. We realized there was no intelligent tool that could detect gambling behavior in real-time and support users proactively. That’s when we built QuitBet — an AI-powered assistant to help people take back control of their lives.
What it does
QuitBet helps users quit gambling by:
- Connecting to their bank account securely via Plaid
- Detecting gambling-related transactions
- Analyzing behavior patterns using Gemini AI
- Providing personalized AI-generated insights and motivational check-ins
- Tracking clean streaks and estimating money saved
- Offering optional accountability features
- Allowing users to reflect with AI-generated surveys
How we built it
We built QuitBet using:
- Flask (Python) for the backend server
- Plaid API to securely access bank transaction data
- Gemini (Google Generative AI) to analyze user behavior and generate insights
- HTML/CSS + Jinja2 for the frontend interface
- Render for deployment
- JSON files for lightweight user session and data storage
We also used the Gemini API to dynamically generate personal survey questions and motivational content based on real financial data and behavior patterns.
Challenges we ran into
- Integrating Plaid’s sandbox with realistic and varied transaction data was tricky
- Parsing transaction data to reliably detect gambling-related activity
- Connecting Gemini AI into our app and formatting prompts effectively
- Handling JSON serialization for complex objects like dates and custom models
- Designing a smooth user experience while handling sensitive topics like addiction
Accomplishments that we're proud of
- Successfully detecting and analyzing gambling transactions from real-time bank data
- Seamlessly combining Plaid and Gemini into a working, full-stack AI tool
- Generating personalized, context-aware check-ins and surveys using AI
- Creating a clean, emotional, and stigma-free experience for users
- Deploying a working version that anyone can test online
What we learned
- How to integrate financial APIs like Plaid and work with transaction data
- How to use Google’s Gemini models to generate high-quality, contextual outputs
- Designing with empathy — especially around sensitive topics like addiction
- Building full-stack Flask apps and deploying them using Render
- Creating clean UX flows for behavior tracking and user reflection
What's next for QuitBet
- Build out a secure database to support more users at scale
- Add SMS or email-based daily check-ins and accountability features
- Train a fine-tuned behavior detection model for more accurate classification
- Offer dashboard visualizations of progress over time
- Partner with health organizations to make QuitBet more accessible to those in need
- Launch a mobile version for broader reach and usability
Log in or sign up for Devpost to join the conversation.