A coding agent on your laptop stops the moment you close the lid. Which is fine for a quick edit — and quietly frustrating the moment you want it to grind through a long refactor, churn a test suite, or keep going while you go do something else. Move it to a server and that limitation just disappears: it works while you sleep, survives your wifi dropping, and hands you the result when you're back. This covers which VPS actually fits a coding agent — Claude Code, Cursor's CLI, Cline, OpenCode or aider — what to pay, and how to keep it running, with no KYC and crypto payment.

What "on a server" actually buys you

• It doesn't stop. A long task runs to completion regardless of your laptop.
• Clean room. The agent gets a reproducible environment; your machine stays free of build junk and half-installed dependencies.
• Isolation. An agent with shell access is powerful. Far better it has that power on a throwaway box than next to your photos and SSH keys.

That last point is underrated: a VPS is the safer home for an autonomous agent, precisely because it's separate.

Sizing it right

The agent itself is light — it calls a model over an API, so the thinking happens elsewhere (the same reason most AI agents need little RAM). What actually consumes resources is your project's toolchain: a big build, a heavy test run, a local database. So size for the project, not the agent:

• 2 GB / 2 cores — comfortable for the agent + a git checkout + light builds.
• 4 GB — a safe default for real projects with proper test suites.
• More only if your build is genuinely hungry or you run several agents at once.

Which plan fits — and how you pay

For a single coding agent on a real project, the sweet spot is 4 vCPU / 4 GB — room for the agent plus your build and test suite. On EQVPS that's the AI-Agent plan ($10/mo); a 2 GB box is fine for light repos, scale up only if your tests are heavy or you run several agents at once.

Two practical calls:

• NAT, not dedicated IP. A coding agent only makes outbound calls — model API, git, package registries — so a NAT plan with port-forwarded SSH is all it needs, and it's cheaper. Reach for a dedicated-IP plan only if you're also hosting something others connect to (a webhook receiver, a preview server).
• No-KYC, paid in crypto. Sign up with an email and pay in USDC or USDT on Base or Ethereum — no card, no ID. Useful if you're spinning up throwaway boxes per agent and don't want each one tied to your identity. A card works too, but the on-ramp has a ~$27 minimum, so topping up a small balance once is smoother for cheap plans.

It's CPU-only and runs from a single datacenter in Germany — fine for a coding agent (the heavy thinking is on the model provider's side), worth knowing if you specifically need a GPU or another region.

Getting it running

SSH in, set up your toolchain and the agent (Claude Code, Cline's CLI, aider — whichever you use), clone your repo, and export your API key:

export ANTHROPIC_API_KEY=...    # or your provider's key
cd ~/myproject

Now the key trick — don't run it in a bare SSH session, or it dies when you disconnect. Run it inside tmux:

sudo apt install -y tmux
tmux new -s agent
# inside tmux: start your coding agent, point it at the task
# then detach with: Ctrl-b, then d

Detach, close your laptop, go to lunch. The agent keeps working on the server. Come back and tmux attach -t agent to see what it did. For scheduled, unattended runs (e.g. a nightly maintenance pass), wrap it in a systemd service instead — same pattern as keeping any process alive.

The honest caveats

• Give it a sandbox, not the keys to the kingdom. Use a non-root user, keep it in a git repo (every change reviewable and revertible), and don't put production credentials on the box. Autonomy plus shell access plus carelessness is how accidents happen.
• Watch your token spend. A long-running agent calling a model in a loop costs real API money — set limits and check in. The server is cheap; the inference is what adds up.
• It's not magic. A background agent is great for well-scoped, verifiable tasks. Point it at something vague and unattended and you'll come back to confident nonsense. Scope tightly.

Within those rails, a coding agent on its own server is a genuinely useful pattern — your work continues whether or not you do. Spin up a 4 GB box, drop in tmux and your agent, and let it cook. The agent can even rent the server itself over MCP if you want the whole loop autonomous.

FAQ

What VPS should I use for Claude Code or other coding agents?

A 4 vCPU / 4 GB box is the sweet spot for one agent on a real project (EQVPS's AI-Agent plan); 2 GB is enough for light repos. The agent itself is light — size for your project's build and test toolchain, not the agent.

Can I run coding agents on a VPS without KYC?

Yes. Sign up with an email and pay in crypto (USDC or USDT on Base or Ethereum) — no ID, no card. Handy when you spin up throwaway boxes per agent and don't want each tied to your identity.

NAT or dedicated IP for a coding agent?

NAT is enough — a coding agent only makes outbound calls (model API, git, package registries), so port-forwarded SSH on a NAT plan covers it. Pick a dedicated IP only if you also host something inbound, like a webhook receiver or preview server.

Can an AI agent provision its own VPS?

Yes — EQVPS exposes an MCP server, so an agent can list plans, pay in crypto and spin up its own box programmatically. Details at /docs.

Why run a coding agent on a VPS instead of my laptop?

Persistence. On a server the agent keeps working when your laptop sleeps, closes or loses wifi — a long refactor or test run survives your day. It also keeps heavy installs and build clutter off your machine, and gives the agent a clean, reproducible environment to work in.

How much RAM does a coding agent need on a server?

The agent process itself is light — it calls a model over an API, so 2 GB handles the agent plus a git checkout and editors comfortably. What eats RAM is your project's own build/test toolchain (a big Node or compiler build), so size for that. 4 GB is a safe default for real projects.

How do I keep the agent running after I disconnect SSH?

Start it inside tmux (or screen). You attach, kick off the agent, detach, and close your laptop — it keeps running on the server. Reconnect later with tmux attach to see progress. For unattended scheduled runs, a systemd service or cron is cleaner.

Is it safe to give an autonomous agent shell access on a server?

Treat it like any powerful tool: run it on a dedicated VPS, not next to anything precious, use a non-root user, keep secrets in environment variables, and work in a git repo so every change is reviewable and revertible. The isolation of a separate box is exactly why a VPS is the safer place for this, not your main machine.

Does the agent need a powerful CPU?

Not for the thinking — that happens on the model provider's hardware. CPU matters only for your project's builds and tests. A 2-core box is fine for most repos; bump cores if your test suite is heavy or you run several agents at once.

Originally published at eqvps.com.