use case · Developers running agents on remote boxes
Run agents on a server, drive them from your laptop's browser
Run the Ringlet daemon on a server and access the agent's PTY from your laptop's browser. Sessions exec inside a sandbox and the PTY is proxied over an authenticated WebSocket, rendered with xterm.js in the local web dashboard.
The problem
- You want an agent to run on a beefy server or near your data, not on your laptop.
- SSH-ing in and multiplexing terminals for several agents is clumsy.
- You don't want the remote agent to have shell access back to your laptop.
How Ringlet handles it
Daemon on the server, browser on the laptop
Run `ringlet daemon --stay-alive` on the server. From your laptop, the local dashboard at 127.0.0.1:8765 attaches to the remote PTY over an authenticated WebSocket.
Sandboxed execution
The daemon starts the agent inside a sandbox — bwrap on Linux, sandbox-exec on macOS — so it runs with least privilege. The agent never gets shell access to your laptop.
Localhost-only, bearer-token auth
The dashboard binds to localhost only and requires a bearer token by default. There is no public cloud endpoint to secure.
The commands
# On the server
ringlet daemon --stay-alive
# On your laptop
ringlet profiles run my-project --remote
# → opens the sandboxed agent PTY in the dashboard
# at http://127.0.0.1:8765 (xterm.js)