StreamPay

Inspiration

We were frustrated with how we had to set up accounts with different AI services like OpenAI and HuggingFace, deposit money, and top up manually as our balances decreased. Our solution hopes to address the growing agentic economy by creating a payment system that's more natural for agent-to-agent transactions.

What it does and how we built it

We consolidate all funding into one wallet for each client. Each client agent, linked to its wallet, opens a payment channel and sets the terms of a service contract. As the provider agent completes units of work, payments stream continuously according to those terms. If work stops, payment stops instantly. When the stream terminates (ideally by satisfaction of both parties), the net transaction is uploaded on-chain.

Challenges we ran into

The biggest challenge was designing a system that enables real-time trust without requiring us to inspect or validate the correctness of an agent’s work. We focused instead on cryptographic verification and payment controls that allow clients to stop payment instantly if value stops being delivered.

Accomplishments that we're proud of

Built a working machine-to-machine streaming payment system on XRPL
Successfully integrated both XRP and RLUSD into a unified interface
Demonstrated real-time payment flow tied directly to work progress
Designed a system that stops payment instantly without disputes or chargebacks## What we learned
Enabling client agents to feel secure with cryptographic verification and ease of terminating streams, while eliminating credit risk for provider agents by ensuring payment comes in as they work.