Skip to main content

Getting Started

This guide takes you from a freshly installed platform to chatting with your own model. It assumes the platform is already running; if not, start with Installation.

1. Deploy a model

Open Models in the portal and create a deployment from a Model preset — pick a chat model from the list, give the deployment a name, and confirm. The platform places it on an available GPU and starts serving it.

No GPU at hand? Connect an external model (OpenAI, Azure, …) instead. It appears alongside in-cluster models and behaves the same from here on.

2. Open the chat

Open the built-in Chat from the portal. It's ready to go — nothing to deploy or connect.

3. Pick your model and chat

Chat's model picker lists every chat-capable model you have access to. Select the one you just deployed and start talking to it.

(Only chat models appear here — embedding and reranker models are filtered out.)

Next steps

  • Call your model from your own code over the AI APIs using an API key.
  • Ground answers in your own documents with RAG.