Quickstart: first endpoint in about ten minutes

This walkthrough assumes you have a console login and one machine you are willing to enrol. Use a test machine rather than a production server for your first run, so that uninstalling afterwards is uncomplicated.

Step one: create a client. Go to Clients and add a record. Even if you are an internal IT team with only one organisation to manage, create the client — reporting, reconciliation and access scoping all hang off it, and retrofitting a client onto an estate that was enrolled without one is tedious.

Step two: create a site under that client. A site is where the endpoint physically or logically lives. One site is enough to start; you can split later and reassign endpoints in bulk.

Step three: generate an install token. From Endpoints, choose the enrol action and select the client and site you just created. The platform issues a token with a limited lifetime. The token is what binds the agent to the right place in your structure at first contact — an agent that enrols without one has nowhere to belong.

Step four: install the agent. Download the build matching the target platform and run the install with the token. On Linux and Windows this is a scripted install; on macOS you can use either the script or the signed installer package.

# Linux — run as root
curl -fsSL https://<your-console-host>/api/agents/download/linux -o aegisone-agent
chmod +x aegisone-agent
sudo ./aegisone-agent install --token "<INSTALL_TOKEN>" --server https://<your-console-host>

Step five: confirm it reported in. Return to Endpoints and refresh. Within a minute or two the machine should appear as online, carrying its hostname, operating system and architecture. Open it and check that the hardware inventory has populated — that is the signal that the agent is not just connected but successfully collecting.

Step six: run a patch scan against it. From the endpoint view, trigger a patch scan. This is the fastest way to confirm the command path works in both directions: the platform can dispatch an instruction and the agent can return a result.

At that point you have a working loop. Everything else in the product — grouping, reporting, reconciliation, remote access — builds on this same enrolled-endpoint foundation.