Claude Code on a VPS with tmux vs Podway

Short answer: if you're comfortable running a Linux server, Claude Code on a VPS inside tmux is a legit, cheap way to keep it running 24/7 and reach it from your phone. You pay with your time: an evening to set it up, then the upkeep nobody mentions. A managed pod like Podway makes sense when that upkeep is worth more to you than a few dollars a month.

Here is the DIY path, step by step, what it costs after day one, and how to decide.

Why a server at all

Claude Code's Remote Control lets you steer a session from the Claude app, but the session runs wherever you started it. Anthropic's docs: "your computer has to stay on and the claude process has to keep running." Close the laptop and the phone has nothing to talk to. So you move the process to a machine that doesn't sleep.

The DIY path, step by step

Rent a small Linux VPS from any provider. A small instance is enough for Claude Code itself; size up if your project builds or runs heavy services.
Lock it down. Create a normal user, log in with SSH keys, turn off password login, and allow only SSH through the firewall. Many people also put SSH behind a private network like Tailscale so the box has no open ports at all.
Install Claude Code with the official native installer, which updates itself in the background:

curl -fsSL https://claude.ai/install.sh | bash

Sign in. Run claude, then /login, and finish the browser step. Remote Control needs a claude.ai subscription login; API keys aren't supported.
Start it inside tmux. The docs say it directly: on a remote machine, "start it inside tmux or screen to keep the session running after you disconnect from SSH."

tmux new -s claude
cd ~/your-project
claude remote-control

Accept the one-time prompts. The first run asks you to trust the folder and to enable Remote Control. Run it in a project folder: in your home directory the trust answer is never saved.
Connect your phone. Server mode prints a session URL; press space for a QR code. Detach with Ctrl-b d and close SSH.
Survive reboots. Add a line to crontab -e so the server comes back after a restart:

@reboot tmux new -d -s claude 'cd ~/your-project && ~/.local/bin/claude remote-control'

Run claude remote-control again in the same folder and it brings back the sessions it was serving, for about four hours after it stopped. After that you get a fresh session.

That's a working setup. For one person who likes servers, it's good.

What it costs after day one

The monthly bill is small. The real cost is a second job: keeping the box healthy.

OS updates and reboots. Security patches need applying, and some need a reboot. If the reboot line above isn't there or breaks, nothing comes back and you find out from your phone.
Claude Code updates. The native install updates itself. Usually that's fine. If you want fewer surprises, set autoUpdatesChannel to "stable" in your settings.
Logging in again. Saved logins can expire or go stale, for example after a plan change, when the docs suggest claude auth logout then claude auth login. Either way you SSH in and sign in again.
Keeping Remote Control alive. It reconnects after short network drops, but not forever. Per the docs, if something keeps refusing the connection, server mode "gives up after roughly 10 minutes and the claude remote-control process exits." Someone has to notice and start it again.
Security. Claude runs real commands on the box. If you relax permission prompts so it can work unattended, give it its own user, keep secrets it doesn't need out of its home folder, and don't run it on a server that holds anything else.
Backups. Your work lives on one disk unless you set up backups yourself, and test restoring them.
Knowing it's down. Nothing tells you. You notice when the phone shows nothing.

None of these are hard. They add up to an hour here and there, usually when you'd rather be doing something else.

What a Podway pod changes

A Podway pod is Claude Code on its own always-on cloud computer, so steps 1, 2, 3 and 8 are already done. You open it from the Claude app on your phone or desktop through Remote Control, on your own Claude Pro or Max plan. On Team and Enterprise plans the org Owner has to enable Remote Control first, which is true on a VPS too. Your files get nightly encrypted backups to separate storage, kept as 7 daily and 4 weekly copies.

What it doesn't change: you still need your own Claude plan, and your Claude usage counts toward it the same way. You also give up some control. A VPS is yours to configure any way you like.

Side by side

	Claude Code on a VPS + tmux	Podway pod
Monthly price	A small VPS from any provider	First pod free, no card; then from $4/month
Your time to set up	An evening if you've run servers before	Create a pod, sign in to Claude
Server hardening and OS updates	You	Podway updates the pod's system image (nightly, or when you click Update)
Coming back after a reboot	Your cron or systemd line	Claude Code and Remote Control come back by themselves; your own services via `podway startup`
Claude Code install	You run the installer	Already there
Steer from your phone	Yes, Remote Control	Yes, Remote Control
Backups	You set them up and test them	Nightly, encrypted, separate storage, 7 daily + 4 weekly
Control over the machine	Full: any provider, any OS, any setup	Less: a managed machine
Claude plan	Pro or Max (Team/Enterprise: Owner enables Remote Control)	Same

When the VPS is the better choice

You already run servers and enjoy it.
You want full control: your provider, your OS, your firewall rules.
You need something on the machine a managed pod might not allow.
You'd rather spend an hour a month than any extra money.

When a pod makes sense

You want Claude Code reachable from your phone and don't want to look after a server.
The machine is a means to an end; the project is what you care about.
You'd rather have backups that already exist than a plan to set them up.
You want to try the always-on setup tonight, free, before deciding.

FAQ

Do I really need tmux?

Yes, or screen. If you start claude straight from an SSH session and close it, the process dies and Remote Control goes offline. Anthropic's docs recommend tmux or screen for exactly this.

Can I use an Anthropic API key instead of a Claude plan?

Not with Remote Control. It needs a claude.ai subscription login on a Pro, Max, Team or Enterprise plan. API keys aren't supported.

What happens when the VPS reboots?

The tmux session is gone. Unless you added a start-on-boot line, Claude Code stays down until you SSH in and start it. With one, claude remote-control comes back and restores the sessions it was serving if the restart was within about four hours.

Does it work on a Claude Team plan?

Yes, on a VPS or a pod, once your organization's Owner turns on Remote Control in the Claude Code admin settings. It's off by default for Team and Enterprise.

Is Podway made by Anthropic?

No. Podway is an independent product. It runs the official Claude Code on your own Claude account.

If the DIY path sounds fun, go for it; everything above is all you need. If you'd rather skip the second job, start a free pod.

Last reviewed October 25, 2026, against Anthropic's Claude Code docs. Spot something out of date? Email support@podway.io and we'll fix it.