I put an AI agent on a $7/month VPS with IRC as its transport layer
Get picks like this daily. Today's launch numbers, the top launches, AI and tech news, and a daily opportunity pick, straight to your inbox.
Want to reach these readers? Sponsor an issue.
The stack: two agents on separate boxes. The public one (nullclaw) is a 678 KB Zig binary using ~1 MB RAM, connected to an Ergo IRC server. Visitors talk to it via a gamja web client embedded in my site. The private one (ironclaw) handles email and scheduling, reachable only over Tailscale via Google's A2A protocol.Tiered inference: Haiku 4.5 for conversation (sub-second, cheap), Sonnet 4.6 for tool use (only when needed). Hard cap at $2/day.A2A passthrough: the private-side agent borrows the gateway's own inference pipeline, so there's one API key and one billing relationship regardless of who initiated the request.You can talk to nully at https://georgelarson.me/chat/ or connect with any IRC client to irc.georgelarson.me:6697 (TLS), channel #lobby.
HN · launched Mar 26, 2026 · 340 upvotes (top 1%) · 0 views
Niche: ai chatbots and conversational agents · 372 launches, 28 in the last 7 days ▼ -33% · Explore this niche →
Category and tags
- Horizontal
- Agent / copilot
- Developer
- AI-native
- Hobby / open-source project
Visit website ↗ Source listing ↗
More like this: all 372 launches in ai chatbots and conversational agents →
Could you build this?
Yes The system connects standard open-source tools (an Ergo IRC daemon, a web IRC client, and an agent script invoking Anthropic's API to inspect Git repos and run commands on a VPS).
Discussion
20 comments analyzed.
Competitors mentioned: Slack as transport layer for multi-agent bots, FastAPI + SQLite message board for agent communication, HTTP polling + acknowledgment-based dedup systems
Concerns raised: IRC single-threaded concurrency handling with multiple users, Message loss when bouncer disconnects or during netsplit, Compromised agent could access email and personal context beyond stated blast radius, Public lobby exposes user questions and potential illegal content hosting risk, Unattended security updates themselves pose security risks (e.g., litellm compromises)
Feature requests: User ignore/block functionality to prevent unwanted command escalation, Request queue with position announcement for concurrent users, Spend attribution logging per session across different model tiers, Crash recovery and message dedup persistence mechanisms
Competitors
Other products that read as similar to this one : 688 launches clear the similarity bar, closest 8 shown.
Attention rank: #19 of 689 (itself plus its competitors, highest first, normalized so YC and Product Hunt are compared fairly).
Launched 147 days after the earliest competitor.
- AgentKeyGate · ph · 2026-11-12 · 1 upvotes · similarity 0.51
- E2a · hn · 2026-05-11 · 48 upvotes · similarity 0.48
- AI agents run my one-person company on Gemini's free tier · hn · 2026-03-08 · 16 upvotes · similarity 0.47
- AgentCheck Free Mini-Check · ph · 2026-11-04 · 1 upvotes · similarity 0.46
- PrivateClaw · hn · 2026-04-24 · 6 upvotes · similarity 0.45
- Diali.ai · ph · 2026-10-28 · 1 upvotes · similarity 0.44
- AgentBus · hn · 2026-03-04 · 10 upvotes · similarity 0.44
- AMA2, messenger built for AI agent · hn · 2026-06-30 · 5 upvotes · similarity 0.44
The niche, labels, vibe code rating and competitor comparison on this page are analysis by Nicheloom. If you use them, please credit nicheloom.com. How you may use our data.