InspirationFreelancers often finish the work before they receive the money. When a client pays late, it can be difficult to tell which payments are still outstanding—and whether current savings can cover essential expenses in the meantime. An unpaid invoice may be expected income, but it is not cash the freelancer can spend today.
PayPilot is designed to help with that uncertainty. Freelancers can review customer, payment and invoice history, then use the Income Shock Lens to explore how their own reserve and monthly essentials affect their estimated cash coverage. The tool uses the numbers the user enters; itdoes not predict future income or count unpaid requests as available cash.
PayPilot’s AI assistant can answer questions using the user’s payment records and prepare a payment request or invoice. It prepares a draft, not an automatic payment: the user reviews and confirms the details before proceeding. PayPal remains the source of truth for payment status.
We built PayPilot with a React and Vite frontend and a FastAPI backend. PayPal Orders v2 supports the payment flow, while Gemini function calling lets the assistant use a limited set of payment lookup and draft-preparation tools. A labeled demo mode lets people explore the app without payment credentials.
Built With
- and
- fastapi
- function
- gemini
- paypal-orders-v2
- react
- sqlalchemy
- sqlite/postgresql
- tailwind-css
- vite
