CashPact

Inspiration

We’ve all set big goals, felt motivated, and then quietly drifted back to old habits, impending deadlines, and burn out. Motivation fades. Consequences don’t. We wanted to build a system where execution is made easy and accountability unavoidable. That’s how CashPact was born.

What it does

CashPact uses the Gemini API to take your goals whether broad, ambitious, or highly specific and break them down into actionable steps you can complete that day. Don’t like the steps? Regenerate them. Finish a task? Mark it done. At the end of the day, your progress is evaluated (prototyped through our “Evaluate” feature for demo purposes). If you fall short, a proportional amount of the money you staked transfers to your accountability partner based on how much you didn’t complete.

How we built it

Node.js backend to manage contracts, evaluations, and logic Supabase for database management, user profiles, and account balance tracking React frontend to create a clean, intuitive accountability dashboard Gemini API to decompose long-term goals into structured, daily action steps

Challenges we ran into

One of our biggest challenges was merging the frontend and backend into a fully functioning system. While each side worked independently, connecting them exposed issues with routing, endpoint mismatches, contract evaluation logic, and database policies. Ensuring data flowed correctly between the React frontend, Node backend, and Supabase database required careful debugging and restructuring.

Another major challenge was working with Gemini’s API. Getting consistent, structured outputs from the AI for goal decomposition required prompt refinement and error handling. We had to handle unpredictable formatting, parse responses reliably, and build safeguards in case the API failed or returned malformed JSON.

Accomplishments that we're proud of

Designing a product which solves a universal pain point Automating contract evaluation and balance transfers Integrating AI-generated goal decomposition into a real execution engine Building a product that combines behavioral psychology, social accountability, and fintech infrastructure

What we learned

On the technical side, we gained experience building secure backend logic, handling database policies, and designing around real-world constraints. On the creative side and non technical side we learned how to ideate at the intersection of multiple pathways, how to plan and prioritize a large project under intense time constraints. And we attended workshops where we gained skills in github, react, ai integration, and hackathon strategies.