Skip to content
INFRASTRUCTURE

Cloud development environments for AI coding agents

Every project runs in a persistent, isolated container with agents already installed. Start work, step away, and reconnect to the same session.

12.4k GitHub stars1.7k forks

Persistent containers with agents already installed

Each CloudCLI project has its own container for the repo, shell, files, Git state, and agent session. It stays available when you reconnect.

Persistent container per project

It does not stop when your laptop sleeps.

Per-developer isolation

Your environment is yours alone, not a shared box.

Agents ready in the environment

Claude Code, Cursor CLI, Codex, and Gemini CLI are ready. Bring your own keys.

Access from every surface

Phone, browser, VS Code, or SSH. Same session, same workspace.

Create a cloud dev environment from a repo

Connect a repo, pick a tier, and land in a running container with the workspace, shell, and agent surface ready.

Create the environment from a repo

Connect a repository, pick the Starter tier, and create a persistent workspace without SSH tunnels, tmux setup, or local machine babysitting.

Open the running workspace

The running environment brings the file tree, terminal, and Claude Code agent panel into one persistent container. The session stays available when you reconnect from another device.

Why agents need cloud development environments

AI coding agents need the same things developers need: a repo, shell, files, credentials, and terminal state that survive device changes and long-running work.

Real compute

Starter gives each environment 2 CPU cores and 4 GB RAM for EUR 7/month. The workspace runs as a container, not a toy sandbox.

Long jobs keep running

Kick off tests, refactors, or migrations, close the laptop, and reconnect later to the same running workspace.

Code and keys stay isolated

Each developer gets a separate Docker container. Credentials are encrypted and forwarded to agents without becoming shared team state.

FAQ

A cloud development environment (CDE) is a containerized workspace that runs on remote infrastructure instead of your local machine. With CloudCLI, that container has your AI coding agents pre-installed, a file explorer, a Git UI, and a shell, and it persists between sessions.

A local environment lives on your machine and stops when the machine sleeps or changes. A CloudCLI environment runs in the cloud, so it survives a closed laptop, a dropped network, and a device switch.

Yes. The agent runs inside the cloud container, not on your laptop. Close the lid and the agent keeps working. Long jobs finish while you are away, and the result is waiting when you reconnect.

Yes. Every developer gets their own Docker container with encrypted credentials. Environments are isolated from each other even when a team shares MCP servers and configuration.

Start a cloud environment for your agents

Create a persistent cloud dev environment with the workspace, shell, files, and agent session ready to reconnect.

No pitch, just answers →