I've been working on CloudRouter, a skill + CLI that gives coding agents like Claude Code and Codex the ability to start cloud VMs and GPUs.

When an agent writes code, it usually needs to start a dev server, run tests, open a browser to verify its work. Today that all happens on your local machine. This works fine for a single task, but the agent is sharing your computer: your ports, RAM, screen. If you run multiple agents in parallel, it gets a bit chaotic. Docker helps with isolation, but it still uses your machine's resources, and doesn't give the agent a browser, a desktop, or a GPU to close the loop properly. The agent could handle all of this on its own if it had a primitive for starting VMs.

CloudRouter is that primitive — a skill that gives the agent its own machines. The agent can start a VM from your local project directory, upload the project files, run commands on the VM, and tear it down when it's done. If it needs a GPU, it can request one.

  cloudrouter start ./my-project
  cloudrouter start --gpu B200 ./my-project
  cloudrouter ssh cr_abc123 "npm install && npm run dev"

  cloudrouter browser open cr_abc123 "http://localhost:3000"
  cloudrouter browser snapshot -i cr_abc123
  # → @e1 [link] Home  @e2 [link] Settings  @e3 [button] Sign Out
  cloudrouter browser click cr_abc123 @e2
  cloudrouter browser screenshot cr_abc123 result.png

What surprised me is how this inverted my workflow. Most cloud dev tooling starts from cloud (background agents, remote SSH, etc) to local for testing. But CloudRouter keeps your agents local and pushes the agent's work to the cloud. The agent does the same things it would do locally — running dev servers, operating browsers — but now on a VM. As I stopped watching agents work and worrying about local constraints, I started to run more tasks in parallel.

The GPU side is the part I'm most curious to see develop. Today if you want a coding agent to help with anything involving training or inference, there's a manual step where you go provision a machine. With CloudRouter the agent can just spin up a GPU sandbox, run the workload, and clean it up when it's done. Some of my friends have been using it to have agents run small experiments in parallel, but my ears are open to other use cases.

Would love your feedback and ideas. CloudRouter lives under packages/cloudrouter of our monorepo https://github.com/manaflow-ai/manaflow.

[1] https://github.com/vercel-labs/agent-browser

Ah, just one step closer to a model with it's own weights file can bootstrap and run itself.

This is cool! I tried it out, running outside my agent with `cloudrouter start .` and got a password request to auth into the server. Opened an issue[1].

[1] https://github.com/manaflow-ai/manaflow/issues/1711

Freaking wow.

It's a cool idea, but personally I don't like the implementation. I usually don't use monolithic tools that cram a lot of different solutions into one thing. For one thing, especially if they're compiled, it's very hard to just modify them to do one extra thing I need without getting into a long development cycle. For two, they are usually inflexible, restricting what I can do. Third, they often aren't very composeable. Fourth, often they aren't easily pluggable/extensible.

I much prefer independent, loosely coupled, highly cohesive, composeable, extensible tools. It's not a very "programmery" solution, but it makes it easier as a user to fix things, extend things, combine things, etc.

The Docker template you have bundles a ton of apps into one container. This is problematic as it creates a big support burden, build burden, and compatibility burden. Docker works better when you make individual containers of a single app, and run them separately, and connect them with tcp, sockets, or volumes. Then the user can swap them out, add new ones, remove unneded ones, etc, and they can use an official upstream project. Docker-in-docker with a custom docker network works pretty well, and the host is still accessible if needed.

As a nit-pick: your auth code has browser-handling logic. This is low cohesion, a sign of problems to come. And in your rsync code:

   sshCmd := fmt.Sprintf("ssh -o StrictHostKeyChecking=no -o UserKnownHostsFile=/dev/null -o ProxyCommand=%q", proxyCmd)

What stops just mentioning AWS/Azure/GCP CLI tools to agents?

Nice, we build something similar at dstack

We recently also added support for agents: https://skills.sh/dstackai/dstack/dstack

Our approach though is more tide-case agnostic and in the direction of brining full-fledged container orchestration converting from development to training and inference

Awesome demo!!!

Thanks for such an enjoyable read!