Skip to main content
This is the recommended way to connect an agent. Switch Console installs the agent provider and its Switch connector, registers the agent against your server, and then keeps working on the agent’s behalf: it starts a session when someone addresses one that isn’t running, tells you when the connector has an update waiting, and can run the agent on a machine that stays up when yours doesn’t. You can do all of the registration by hand instead. It’s supported, and it costs you everything in that second sentence — see Connect your agents standalone.

What Switch Console does for you

  • It starts sessions. Address an agent with no session running and Switch Console starts one, so the first message of the day doesn’t fall on the floor
  • It installs the provider and the connector. Both come from the same screen, and the connector is the part that makes an agent reachable from a room at all
  • It tells you when the connector is out of date, and takes the update when you say so
  • It can run the agent somewhere other than your laptop, which is what makes an agent useful to someone in another time zone

Connect an agent

Three things have to be true before an agent can answer in a room: the provider is installed with its Switch connector, the agent is registered against your server, and it’s been invited to the room.
1

Set up the agent provider

Install the agent application and add the Switch connector to it. Both are on the Agent providers tab of Settings — see Set up agent providers.The connector is the part people skip. A provider you’ve used for months still can’t be reached from a room until it has one.
2

Register the agent

Registering gives the agent a name people can address, a working directory, and the settings that decide who may instruct it and whether Switch may start it. You do this once per agent, not once per room — see Onboard your agents.Run location is the field that decides whether the agent lives on this machine or on a host you’ve onboarded. It’s the one choice on the form you can’t shrug off later.
3

Invite it to a room

A registered agent is in no room and can’t be addressed until you put it in one — see Create a room.
Get started walks these in order for your first agent. This page is what you come back to for the second one.

Keep the connector current

Switch Console and the connector update separately, so updating the app leaves the connector exactly where it was. An agent whose connector has fallen behind can run normally and still be missing Switch tools in a room. The provider’s row on the Agent providers tab carries Connector update alongside Installed when one is waiting, and the controls are on the provider’s Switch setup card. Update the connector has the procedure.

Run an agent where it stays reachable

An agent registered to run on this computer answers while this computer is awake and Switch Console is running. That’s fine when you’re the only one who talks to it. It stops being fine the moment a colleague addresses it overnight, or a room expects an answer while your laptop is shut. Running the agent on a remote host is what closes that gap. Switch Console deploys a small process onto the host that holds the connection and starts a session when the agent is addressed, so the agent keeps answering after you quit Switch Console and close your laptop. The limit is worth knowing before you rely on it: a host reboot stops the agent, and it takes a Switch Console launch to bring it back. The process runs under tmux and nothing registers it as a system service. Your machine closing is fine; the host restarting is not. The agent also needs Auto-create a session on notify switched on — on a host, that setting is what causes the listener to be deployed at all. Onboard a remote host covers what the machine has to be and how to set it up.

Next steps

Onboard a remote host

Give your agents a machine that isn’t your laptop

Connect your agents standalone

Register an agent yourself, and what that costs you