Your first 15 minutes
By the end of this you will have Rookery running and one agent that actually does something.
1. Install it
Section titled “1. Install it”On Linux and macOS:
curl -fsSL https://rookery.cloud/install.sh | shOn Windows, in PowerShell:
irm https://rookery.cloud/install.ps1 | iexPrefer a container, or want to read the script first? See Docker and Linux server.
2. Set it up
Section titled “2. Set it up”rookery onboardOne interactive pass does the whole setup: it resolves the two keys and explains which one matters, creates your owner account — there is exactly one per installation — offers any missing host tools, reports the coder situation, and on Linux installs and enables the service so Rookery survives a reboot.
Anything it skips is repeated in a closing Still to do list, so a partial setup never looks like a finished one.
3. Start it
Section titled “3. Start it”On Linux, onboard already started the service. Otherwise:
rookery serveIt listens on 0.0.0.0:8080 by default. The database is created and brought up
to date automatically on start — there is no separate migration step.
Open http://localhost:8080 and log in.
4. Create a workspace
Section titled “4. Create a workspace”A workspace is a sealed world: its own knowledge, credentials, connections and agents. You will be asked for a master password for it, and asked again every time you enter it.
5. Give it a model to think with
Section titled “5. Give it a model to think with”Rookery needs a model. You can point it at a coder tool you already have installed, at a hosted provider, or at a model running on your own hardware.
See Choosing a model.
6. Build your first agent
Section titled “6. Build your first agent”Go to Agents → New, and describe what you want in your own words. Something small and real works best:
Every morning, check whether rookery.cloud is reachable and tell me only if it isn’t.
Rookery will ask a couple of questions, then propose a plan and offer
Approve & build (typing approve does the same). The button appears only
once there is a plan — while it is still asking questions, the answer it wants
is yours, not a click. Then it builds the agent, runs it for real, and shows you
what happened before saving anything.
Where to go next
Section titled “Where to go next”- Workspaces — what is isolated, and what that means.
- Linux server — keeping it running after you close the terminal.