Codex in the cloud
with a machine of its own
Run the Codex CLI on a machine that stays. Your repository, its tools and the conversation are there every time you come back.
No newsletter. One email when your access is ready.
What it gets you
A repository that stays
Clone once. The code, the dependencies and the build cache stay on the agent's disk.
OpenAI or OpenRouter
Use an OpenAI key, or an OpenRouter key to reach other models such as DeepSeek or Kimi.
From your terminal
waken deploy --harness codex sends your project folder and starts the agent.
Answers anywhere
Write to it from the dashboard, the CLI, Telegram, Discord or Slack.
Rewind
A checkpoint every 15 minutes, and one click to undo a bad change to the whole machine.
Spend cap
Set a budget per agent. Codex stops cleanly when it reaches it.
Leave it a task.
Find it done.
Hand Codex a task from your project folder and close the laptop. It works in its own copy of the repository, and when you are back, the changes are waiting for you to review.
Questions
and answers.
Keep
reading.
- Run Claude Code in the cloud, 24/7Host Claude Code on its own cloud machine. It keeps its files and memory, sleeps when idle, wakes in 0.08 s, and answers from Telegram, Slack or your terminal.
- OpenClaw hosting: run OpenClaw 24/7Host the open source OpenClaw agent on its own cloud machine. Persistent memory, any model on your own key, sleep when idle, checkpoints and rewind.
- Running AI agents on Waken vs on your laptopAgents on your laptop stop when the lid closes and can touch all your files. Here is how a Waken agent compares, with its own machine in the cloud.
- A secure sandbox for every AI agentRun each AI agent in its own virtual machine, with checkpoints every 15 minutes and one-click rewind. Encrypted at rest, hosted in Europe.