Art of the Smart

Thoughts and lessons from a techie, entrepreneur, and curious observer of the world.

Rapid Local Agentic Setup

Here's the quickest way to set up your local agentic LLM.

  1. Download and install Ollama, a known-safe local llm runner. Upgrade to llama.cpp for a potential speed boost later.
  2. Run one of the following commands depending on your platform. It will generate a Google Gemini query asking it to select the best Ollama model for your hardware.
    1. Linux, Mac Run terminal and paste this in and run it:
    2. Windows Run PowerShell and paste this in and run it:
  3. Paste the result of that command into Gemini and it'll tell you which Ollama model is probably best for your hardware.
  4. Select fast, balanced, or performance options (or download all 3) in Ollama and get them downloaded and installed.
  5. From the Apps pane in Ollama, select the Command Line harness of your choice (e.g. Copilot or Claude Code).
  6. Open Powershell or Terminal and navigate to the directory of your project. Paste in the Ollama command it put in your clipboard.

Done!