I put an AI agent on a $7/month VPS with IRC as its transport layer
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…
In plain words
This project runs two AI agents on budget infrastructure: a public conversational agent (nullclaw) accessible via IRC and an embedded web client, and a private agent (ironclaw) handling email and scheduling over Tailscale. Built as a 678 KB Zig binary using minimal RAM on a $7/month VPS, it uses tiered inference with cheaper Haiku models for chat and more capable Sonnet models only when needed, maintaining a hard cost cap of $2 per day. The system consolidates API billing through a single gateway, making it an economical approach to running autonomous agents.
written from the facts on this page · September 2026
From the sources
In the maker’s words, at launch
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.
Does the same job
all alternatives →
- NCNanoClaw – “Clawdbot” in 500 lines of TS with Apple container isolationFeb 2026 · github.com · ▲533
I’ve been running Clawdbot for the last couple weeks and have genuinely found it useful but running it scares the crap out of me. OpenClaw has 52+ modules and runs agents with near-unlimited permissions in a single Node process. NanoClaw is ~500 lines of core code, agents run in actual Apple containers with filesystem isolation. Each chat gets its own sandboxed context. This is not a swiss army knife. It’s built to match my exact needs. Fork it and make it yours.
- AAAI agents run my one-person company on Gemini's free tier – $0/monthMar 2026 · ▲16
I'm a solo dev in Taiwan. I built 4 AI agents that handle content, sales leads, security scanning, and ops for my tech agency — all on Gemini 2.5 Flash free tier (1,500 req/day). I use ~105. Monthly LLM cost: $0. Architecture: 4 agents on OpenClaw (open source), running on WSL2 at home with 25 systemd timers. What they do every day: - Generate 8 social posts across platforms (quality-gated: generate → self-review → rewrite if score < 7/10) - Engage with community posts and auto-reply to comments (context-aware, max 2 rounds) - Research via RSS + HN API + Jina Reader → feed…
More ai this month
the category →
I trained a 125M-parameter transformer to autocomplete piano performances in real time (~108 notes/sec on an iPhone 15). The idea is basically GitHub Copilot or Tabnine, except instead of prompting it with code, you prompt it by playing a few notes on a MIDI piano. The model then continues what you played, entirely on-device. The app is free if anyone wants to try it. Happy to answer questions about the model, training, Core ML, or the many things that didn't work.
AI · 17d ago · simedw.com
Astute▲585Automate your B2B brand going viral, with new media creators
AI · 18d ago · company-app.joinastute.com
Hey HN, Henry from Cactus here! We previously released Cactus Needle, a 14MB agentic LLM for tool call, device use, and structured extraction for phones, wearables, smart homes, small robots and microcontrollers. We got really great feedback here, and have now incorporated the suggestions to release Needle 2. The whole model is a single 14MB binary that runs a full session in 28MB of RAM; 45m parameters at 2bit compression. Needle hits 500 tokens/sec decode speed on a Raspberry Pi 5, sits between 400-1,500 tokens/sec on VR devices like Meta Quest 3S and Apple Vision Pro, and ranges…
AI · 27d ago · cactuscompute.com
Launched alongside, March 2026
the whole month →
Switch from ChatGPT to Claude with import memory feature
AI · Mar 2026 · claude.com
