🎯 Inspiration Splitting bills with friends usually feels stressful or awkward. Apps today make it formal and boring. We wanted to create something different — KillBill — a fun, AI-powered bill-splitting experience where you just snap a photo, say a few words, and let KillBill handle the messy math. Plus, it roasts your spending habits and turns managing money with friends into something you'd actually enjoy.
⚡ What it does Upload a photo of your bill or receipt.
KillBill automatically parses the bill and extracts items and prices.
You simply tell KillBill who had what (“I ate the pizza, we all had drinks”) and it calculates fair splits.
A smart agent suggests smarter splits based on past habits (e.g., “Alex doesn’t drink — skip beers for him?”).
A fun analytics tab shows weekly spend summaries, jokes, and playful roasts about your habits.
🛠️ How we built it Frontend: Mobile-first web app using React/Next.js for a clean, simple, and fun experience.
Backend: Python FastAPI server to handle image uploads, OCR parsing (with Tesseract), and lightweight NLP for user instructions.
AI Agents: Fetch.ai Autonomous Economic Agent suggesting optimized, fair bill splits based on user history and preferences.
Analytics Module: Tracks weekly expenses and generates funny, personalized insights based on user behavior.
🚧 Challenges we ran into Making OCR parsing reliable for messy, real-world receipts.
Designing a simple yet flexible natural-language command system ("Split pizza for me, drinks for everyone").
Seamlessly integrating an intelligent Fetch.ai agent within fast hackathon deadlines.
Balancing fun UX with the accuracy needed for real money splitting.
🏆 Accomplishments that we're proud of Built a full flow: photo ➔ parsing ➔ smart splitting ➔ analytics, all within a few hours.
Successfully designed and deployed a Fetch.ai agent that personalizes the splitting experience.
Delivered a product that feels fun and non-stressful even when dealing with finances.
Maintained modular, scalable code architecture ready for future extensions.
📚 What we learned How to combine OCR, NLP, and agent-based AI into a seamless app experience.
How even lightweight personalization can dramatically improve user delight.
How important it is to make financial tools emotionally positive — not cold or boring.
How humor and friendliness can be core features, not just decoration.
🚀 What's next for KillBill Smarter natural language processing for messier or more casual instructions.
Build richer user profiles that recognize more spending patterns and optimize splits over time.
Integrate real-time payment links (Venmo, PayPal) to settle up instantly.
Gamify the experience with badges, challenges, and rewards.
Built With
- aeas
- cloud
- css
- fastapi
- fetch.ai
- next.js
- ocr
- postgresql
- python
- python)
- react.js
- storage
- tailwind
- tesseract
Log in or sign up for Devpost to join the conversation.