Give Claude Code an always-on box — and drive it from your phone.

Claude Code is great, but it lives in a terminal on the machine in front of you. Kick off a long run and you're chained to that laptop; close it and the session goes with it. This puts Claude Code on a box that stays on, in your own cloud, and hands you a phone and desktop client for it.

The problem with a terminal-bound agent

An agent running in your local terminal is only alive while your machine and that session are. Long jobs — a big migration, a full test suite, a multi-file refactor — are exactly the ones that outlast your seat at the desk, and they're the ones that get cut off when you close the lid or lose Wi-Fi. You end up babysitting a window instead of walking away.

Put it on a box, keep your own account

Wetlether runs Claude Code on a small always-on server in your own cloud, signed into your own Claude account. The app is a thin client:

Same Claude Code you know, unhooked from a single terminal. The session belongs to a box you own, not to the laptop you happened to start it on.

And it's not only Claude

If you want, connect GPT, Gemini, Grok, or Perplexity on your own keys too, and put several models on one hard question — they debate it and hand you the answer that survived, instead of you running the same prompt through four tools by hand.

The honest tradeoff. You run the box yourself (a few dollars a month on Railway, Render, or similar) and bring your own Claude account — more setup than a hosted app, on purpose, because it keeps the agent, the keys, and the code yours. Free tier to start; paid tiers add the multi-model crew. We never meter you.

See how it works →

Wetlether — Claude Code on a box you own, driven from anywhere. Free to start.

Keep reading