How I Built an Autonomous AI Agent That Earns USDC While I Sleep

Description and use

An autonomous agent that accepts HTTP requests, verifies x402 payments, processes tasks, and earns USDC micropayments, running on a VPS.

Problem

Developers need a way to run autonomous agents that perform useful tasks and receive micropayments without continuous human supervision.

Outcome

The agent processes requests in under 120 s, achieves a 99.5% success rate, and collects USDC micropayments per call, demonstrating a self‑sustaining service model.

Workflow

• Accept HTTP request via aiohttp
• Extract and verify x402 payment headers with x402-py
• Parse JSON payload
• Enqueue job into Redis Queue (RQ)
• Planner selects tool based on payload
• Executor runs tool and returns result
• Observer emits Prometheus metrics
• Respond to client with job ID

Evidence

All components run in the same Python process for simplicity; they could be split into separate containers or VMs if higher isolation is needed.

Source

The agent’s main entry point is an asyncio task that runs forever.

Source