Complete first launch and choose a model

Before you start

Install the issued desktop build. These steps describe desktop IAM.Agent; agent.iamrouter.ru is a separate web entry.

Click path

1. Click Next on the welcome screen.
2. In Choose your engine, select Provider. Local Ollama must already run at localhost:11434; an OpenAI-compatible provider needs its endpoint.
3. Set Default model to an available model. The shown model is a sample, not an installation action.
4. Click Next and select the working directory or Skip it for now.
5. Review Tools and confirmations and shortcuts, using Next.
6. Review All set and click Enter the agent.
7. Use Settings to change configuration later; verify provider connectivity before real work.

Enlarge screenshot

1 2

1. Skip
2. Next

Verify and recover

Opening the workspace does not prove inference connectivity. Send a harmless model request to verify it. Check local server/model or cloud settings after connection failures. Native file and shell operations require desktop runtime.