IGEL OS can now run AI models locally on endpoints

IGEL announced that IGEL OS can run language models locally using Ollama, revealed at Now & Next 2026 in Miami.

The AI Armor feature adds runtime security and tracks NPU/GPU access while allowing admins to block unapproved processes.

UMS will include a central policy engine with pre-built compliance templates and controls for who can use Ollama and which prompts are allowed.

Local models like Google’s Gemma 3 can generate responses on-device without cloud tokens or leaving the corporate network.