diff --git a/content/4.agent-mode/0.agent-mode.md b/content/4.agent-mode/0.agent-mode.md
index afe27ac..20447d5 100644
--- a/content/4.agent-mode/0.agent-mode.md
+++ b/content/4.agent-mode/0.agent-mode.md
@@ -116,28 +116,6 @@ The onboarding screen also includes basic terminal guidance:
 - **Windows:** Open **Windows Terminal** or **PowerShell** from the Start menu.
 - **Linux:** Open your preferred **Terminal** app from the applications menu.
 
-## Use Local AI Models
-
-Agent Mode can also use local inference servers as backends. Local providers are discovered from providers you have already configured in the Model Hub with a local-compatible inference endpoint, and each enabled provider appears as its own agent in the Agent selector.
-
-Examples include local or self-hosted providers such as Ollama-compatible servers, llama.cpp, MLX, and other OpenAI-compatible local inference servers.
-
-This setup currently requires **Claude Code CLI** to be installed. Local backends are also generally not ideal for every agent CLI workflow and usually work best when you have a strong local model and enough system resources.
-
-::steps
-### Configure a local provider in Model Hub
-Add or configure your local inference server in the Model Hub so it has a working inference endpoint.
-
-### Open Settings > Agent Mode
-Go to the **Local Providers** section in Agent Mode settings.
-
-### Enable the provider
-Turn on the local providers you want Agent Mode to expose.
-
-### Start or refresh an Agent session
-Enabled local providers appear as separate agents in the Agent selector. Refresh agents if needed.
-::
-
 ## Major Parts
 
 ### Sessions and workspaces