Let your AI agents pay their own way.Safely.
I'm Edward Zhou, a full-stack engineer of 13 years. I help teams design and build on-chain payments and spending controls for AI agents.
SERVICES
What I do
Prototyping
A working prototype of one agent-payment use case in 2–4 weeks.
DeliverableA working prototype deployed to testnet, the full source code and a walkthrough call
Architecture review
What your agent can spend, what it can do, and what happens when it fails.
DeliverableA written report with ranked risks, recommended spending controls and a follow-up call
Technical due diligence
An independent read on a project's contracts and architecture, for investors or teams.
DeliverableA due-diligence report covering findings, red flags and a clear go / no-go view
WORK
An agent wallet that can’t overspend
Live demo · test network
An AI agent pays for API calls on its own through x402. Its wallet is a smart contract: any payment that breaks a rule is refused before money moves, and it pays services outside its list only if they have a good ERC-8004 reputation from reviewers I trust. Running live on the Base Sepolia test network.
- Spending limits enforced in the contract: per payment, per day and per task
- Every payment, refusal and rating is a public on-chain record anyone can verify
- Impostors and unknown services are refused; trust is earned through on-chain reputation
ABOUT
Who you’d be working with
I've built software in Australia for many years, across front end, back end and production operations. Yunshu AI is my independent practice. I take on a small number of technically demanding projects, and work in both English and Chinese.
Stack
- TypeScript
- Rust
- Solidity
- Solana
- NEAR
- AWS
- x402
- ERC-8004
CONTACT
Got an idea? Let's talk for 20 minutes.
No charge, and you'll get short written notes afterwards.
- hello [at] yunshu.ai