We are extending our LLM-driven Kali series, where natural language replaces manual command input. This time however, we are doing everything locally and offline. We are using our own hardware and not relying on any 3rd party services/SaaS.

Kali Linux

A step-by-step guide to running a fully local, offline LLM setup on Kali Linux using Ollama and the 5ire GUI client. Covers installing NVIDIA proprietary drivers with CUDA support, setting up Ollama with models like qwen3:4b and llama3.1:8b, deploying the mcp-kali-server MCP server, and connecting everything through 5ire so natural language prompts can invoke penetration testing tools like nmap — all without any third-party cloud services.

Kali & LLM: Completely local with Ollama & 5ire