# A coding agent, on a machine of its own.

URL: https://munito.ai/cloud-agents

Hand a task to Claude Code, Codex, DeepSeek Harness and five more. Each one runs in its own sandbox on open models in the EU, and sends the patch back.

Say what the work is, and a coding agent does it on a machine of its own. The patch comes back; nothing runs on your laptop.

- Eight harnesses
- A machine in 772 ms
- Public preview
- Per-second rates

### Command line

```bash
npm install -g @munitoai/cli
```

```bash
munito agent dsh "add a retry to the HTTP client"
```

## One command carries the whole task.

Name a harness and say what to do. Munito AI starts a machine, uploads your git worktree, runs the agent, applies the patch it wrote and deletes the machine. In our own measurement the machine was ready in 772 milliseconds.

## Eight harnesses, one key.

Claude Code, Codex, OpenCode, DeepSeek Harness, Hermes, pi, OpenClaw and Prime Agent, each pointed at our open models. One key pays for the machine and for the tokens. Your own Anthropic or OpenAI key works too, and so does your own subscription.

## Read the repository, then change it.

An agent that can run commands can also read the code first. In one measured run, DeepSeek Harness answered a question about a repository in 2 seconds at 251 tokens a second on GLM-5.3-Flash, then wrote four files in 1 minute 54 seconds.

## Watch it work, or do not.

The same machine carries a desktop, so you can open a screen and see the agent type. Leave it out and the run is a command that prints a patch, which is what a queue or a pipeline wants.

## Give it the smallest key you can.

Every program in the machine reads its environment, and so does the agent. Use a key you can revoke, with the permissions and the spend limit of this one task. A subscription sign-in happens inside the machine, over SSH, and never passes through the API.

## Keep the machine when the task is long.

A run ends with its machine by default. Keep it instead, and you have a persistent computer: open a shell over SSH, pause it to hold the disk and the sign-ins, and resume where the agent stopped.

- [Read the Cloud agents guide](https://munito.ai/reference/sandboxes/cloud-agents)
- [Get the command line tool](https://munito.ai/reference/cli)
- [See the prices](https://munito.ai/pricing#sandboxes)

## Common questions

### Can I use Claude Code, Codex, OpenCode or Hermes with open models?

Yes. Claude Code connects to the Anthropic-compatible endpoint, and Codex to the Responses API. OpenCode and Hermes use Munito AI as a custom OpenAI-compatible provider. LangChain and LlamaIndex have guides too.

### Does Munito AI keep my prompts and outputs?

No, not by default. Zero data retention is the default policy: after delivery, Munito AI removes the temporary request content. It keeps content only when a request asks for it, and then for 30 days.

### Where does Munito AI process my requests?

The EU host processes each request in the EU, the US host in the US, and the global host in any region. Each response names its region in the munito-region header.

### What are the rate limits?

Your limits adapt to your usage. A 429 means your traffic is above your limit, and a 503 that the model has no free capacity at that moment. Each response carries your current limits in its headers.

https://docs.munito.ai/faq