Your connection, step by step.
Google and Apple sign-in are awaiting provider configuration. Windows has been tested; native macOS build and execution validation are pending.
01 / Open a workspace
Once sign-in is enabled, continue with Google or Apple on the authentication page. Existing Google-owned email accounts can reconnect to their previous workspace after verification. Apple Hide My Email can represent a different account.
Open sign-in →02 / Register your agent
Get the agent build from your deployment administrator. In Devices → Add device, choose your operating system and issue a one-time enrollment token. On Windows, place commander-agent.exe in a dedicated folder and open PowerShell there.
.\commander-agent.exe enroll --server https://rdc-api.duckdns.org --name My-PC --allow-directory "D:\data" .\commander-agent.exe run
Use the token-entry instructions shown in the device dialog. Keep the agent running. If a configuration already exists, run the agent instead of enrolling again. Do not delete an existing device identity unless you intend to replace it.
03 / Connect your AI client
In a client that supports remote MCP, create a custom connection using the address below and OAuth authentication. Use the OAuth Client ID issued for that client by your Commander administrator. Sign in and review the requested permissions.
https://rdc-api.duckdns.org/mcp
04 / Start with a read-only request
Try: “List my connected computers, then show the files in my allowed project folder.” Enable terminal access separately when needed. Review activity in the dashboard and on your agent.
Open device management →