Show HN: I put an AI agent on a $7/month VPS with IRC as its transport layer

What it is

Picture an AI assistant that doesn't live in Slack or Discord, but in IRC—the text-only chat protocol your tech-savvy uncle used in 1995. The agent sits on a basic virtual private server, joins IRC channels, and responds to commands. Think of it as ChatGPT meets old-school Unix tools: lightweight, scriptable, always-on.

Why it matters

This matters if you're tired of complex cloud setups or vendor lock-in. IRC is simple text over TCP—no APIs to learn, no rate limits, no OAuth dance. You control the entire stack. For personal automation or small team tools, this approach costs less than a sandwich and runs on hardware from 2015. It's proof that boring infrastructure beats bleeding-edge complexity for most use cases.

Key details

• •Uses IRC (Internet Relay Chat), a protocol from 1988 that's still maintained and ubiquitous
• •Runs on a $7/month VPS—likely 1GB RAM, single core, enough for small-scale agents
• •No webhooks, no polling loops—IRC's push model handles real-time events natively
• •Agent called 'Nullclaw Doorman' handles commands in IRC channels like a chatbot
• •Full control over infrastructure: no cloud vendor dependencies or surprise bills