SpareSense

Inspiration

Saving money often feels hard because it requires intentional effort, manual transfers, budgeting discipline, and constant reminders. We noticed that while people spend money every day, they rarely save with the same frequency. At the same time, traditional financial systems aren’t designed to support tiny, frequent savings due to fees and delays.

We were inspired to flip this behavior by making saving automatic, invisible, and effortless. By combining Capital One’s Nessie API, Solana’s blazing fast and low-fee blockchain, and Gemini AI’s ability to generate personalized insights, we set out to make saving both effortless and understandable.

What it does

SpareSense is an AI-powered micro-savings platform that automatically rounds up everyday purchases and invests the spare change into a Solana wallet, while also providing intelligent spending insights.

When a user makes a purchase:

• The amount is rounded up to the nearest dollar.
• The difference is calculated as micro-savings.
• Savings are invested to a Solana wallet.
• Gemini AI analyzes spending patterns and generates personalized advice on how to increase savings, as saving is just as important as investing. Users don’t just see how much they saved. They understand why and how they can improve. This visualization incentivizes them to cultivate better spending habits.

How we built it

SpareSense is built using a clean, scalable backend-first architecture:

• FastAPI backend to handle transaction processing, round-up calculations, savings aggregation, and API orchestration.
• Capital One Nessie API to simulate realistic banking data and purchase transactions.
• Gemini AI to analyze spending patterns and generate personalized insights in plain language.
• Solana SDK to transfer micro-savings on Solana Devnet with near-zero fees.
• React frontend to display savings totals, AI-driven insights, and blockchain confirmations.

Challenges we ran into

One challenge was transforming raw transaction data into insights that were accurate, responsible, and easy to understand. Prompting Gemini AI required careful tuning to ensure advice was helpful without being generic or misleading insights.

Another challenge was handling extremely small monetary values while ensuring precise conversion into lamports for Solana transactions. We also designed the system to guarantee that no sensitive financial data is ever stored on-chain, while still preserving transparency and verifiability.

Accomplishments that we're proud of

• Successfully integrating a FastAPI backend with financial APIs, AI services, and Solana SDK.
• Generating real-time, personalized spending insights using Gemini AI.
• Executing micro-savings transactions on Solana with instant confirmation.
• Building a complete end-to-end fintech product within a 36-hour hackathon. Most importantly, we built a product that actively encourages better financial habits, not just passive tracking.

What we learned

• We learned that AI is most powerful when it explains data rather than overwhelming users. Gemini AI enabled us to turn complex transaction histories into clear, actionable advice.
• We also learned how blockchain can enable new financial behaviors when used intentionally. Solana’s low fees and speed made micro-savings practical in ways traditional financial systems cannot support.
• Finally, we gained hands-on experience designing real-time systems with FastAPI, balancing performance, security, and usability under tight time constraints.

What's next for SpareSense

Next, we plan to:

• Add AI-powered savings goals and progress tracking.
• Introduce predictive insights, such as monthly savings forecasts.
• Support multiple bank accounts and wallets.
• Explore real-world banking integrations and compliance-ready deployments.

Our long-term vision is to build an intelligent savings assistant that works quietly in the background—helping users save smarter without changing how they spend.

Built With

• fastapi
• gemini
• javascript
• python
• react
• solana