Skip to main content

First-run setup

1. Create the administrator​

The first-run path depends on the Engine host:

  • macOS or Windows app: open the app on the host computer. A new 0.8.1 installation requires you to create the first administrator with a name, real email address and password. The administrator is created during first-run setup; 0.8.1 has no separate “set up remote login” step.
  • Linux: use only the one-time setup URL printed by the installer, then create the administrator with a real email address and a password you can store securely.

Complete first-account setup locally before exposing an Engine to remote users. Do not share the Linux setup URL or administrator credentials.

2. Connect a model​

Choose the model connection that matches your account:

  • a supported API provider and API key;
  • ChatGPT subscription connection where offered;
  • a compatible custom or locally hosted model endpoint.

Run the connection test before continuing. A successful network request is not enough: Clawuno must also be able to use the selected model and response format.

Credentials are saved in the deployment you control. If you select a cloud model, prompts and the data you intentionally send are still processed under that provider's terms.

3. Enter the starter Workspace​

After model setup, Clawuno prepares a starter Workspace and an Assistant Agent. The suggested tasks only fill the composer; you decide whether to send them.

Create a small file as your first test, then open Files to confirm the output appears in the Workspace.

4. Meet Uno​

Uno is the system assistant, not the Agent assigned to every task. Open Uno when you want help navigating Clawuno, finding work, configuring an Agent, or understanding the current Workspace.

Try asking:

Show me where to create a new Agent and explain the settings I need for a research assistant.