host ALL your AI locally



AI Summary

Summary: Building a Local AI Server

  • Purpose: Created a local AI server named Terry for personal use and daughters’ education.
  • Features:
    • GUI with chat interface
    • Back chat histories
    • Multiple AI models
    • Stable diffusion integration
    • Added to notes application
  • Hardware:
    • Server named Terry
    • Components include Leon Lee case, ASUS motherboard, AMD Ryzen CPU, 128GB RAM, Leon Lee water cooler, two Nvidia 4090 GPUs, Samsung SSDs, Corsair power supply.
  • Software:
    • Initially struggled with Ubuntu, switched to Pop!_OS
    • Demonstrated setup on a laptop
  • Installation Steps:
    • Install Alama for AI models
    • Use Windows Subsystem for Linux (WSL) for Windows users
    • Install Docker for web UI
    • Install Open Web UI for a feature-rich interface
    • Install Stable Diffusion for local image generation
  • Customization:
    • Restrict models and commands for daughters’ use
    • Integrate AI with Obsidian
  • Demonstration:
    • Showed Terry’s capabilities
    • Showed how to use Open Web UI and Stable Diffusion
    • Integrated AI into Obsidian
  • Privacy and Control:
    • Emphasized local, private control over AI
    • Showed how to restrict AI capabilities for educational purposes

Additional Notes:

  • Obsidian Integration:
    • Used BMO Chatbot plugin to integrate AI into Obsidian
  • Community and Learning:
    • Encouraged joining Discord community for further discussion
    • Mentioned IT Pro for learning Linux and IT skills
  • Documentation:
    • Provided a guide on Network Chuck Academy for the setup process

Conclusion:

The local AI server provides a private, customizable, and powerful AI experience that can be integrated with note-taking applications like Obsidian and controlled for educational purposes. The setup process involves installing various software components and can be done on a personal laptop.

Get the computer build details from the video description