Skip to content

Use cases

Agents where the code is, UI where you are

Helicon ships the same React UI as a desktop app and as a web app pointed at a daemon elsewhere. Put the daemon and the muse CLI on the workstation or server that holds the code, and supervise from a browser, so the agent runs next to the repository while you stay on a laptop.

Last updated · Markdown version

Why not just SSH

SSH gives you a terminal on the far machine, which means you are back to scrollback, no session list and no approval queue. The point of the remote daemon is keeping the interface while moving the execution.

Situations this fits

  • A monorepo whose test suite needs more machine than a laptop has.
  • Code that must stay on a particular machine.
  • Long runs that should not stop when the laptop sleeps.
  • A Linux box for the agent while you work on Windows or macOS.

What to get right

  • Node 22 or newer on the remote machine, since it is running from source.
  • muse login on that machine, since credentials live with the CLI.
  • Your own tunnel, VPN or reverse proxy. Do not expose a daemon with shell access to the open internet.
  • Strict approval modes. Remote execution is exactly where allow all is a bad idea.

Frequently asked questions

Is the web build a cut down version?

No. It is the same React UI. Only the daemon's location differs.

Can several people share one daemon?

Multi user deployment is out of scope today. Helicon targets one developer's own machines.

Does my code pass through helicon.sh?

No. Nothing about your code, prompts, threads or files leaves your machines.

Same Muse Code. Same subscription. Better interface. 

Free and MIT licensed. Signed Windows installer, universal macOS DMG, and a web build against a daemon you run.