The Best Way to Run a Local AI (2025)

Run AI models locally on your PC with LM Studio — no coding, no API costs, complete privacy. This beginner-friendly guide covers installation, finding the right models for your GPU's VRAM, and getting your first local AI chat running in minutes. Products in this video (affiliate links support the channel at no extra cost to you): - The GPU used in this video with 16GB VRAM — ASUS RTX 5060 Ti: https://www.amazon.com/dp/B0F7WB6LSH?tag=hakehardware-20 - Great budget entry point for local AI with 8GB VRAM — MSI RTX 4060 Ventus 2X: https://www.amazon.com/dp/B0C8BPW1SP?tag=hakehardware-20 - Fast NVMe storage for your AI model library — Crucial P3 Plus 2TB: https://www.amazon.com/dp/B0B25ML2FH?tag=hakehardware-20 - Take your AI models on the go with USB 3.2 speeds — Samsung T7 Shield 2TB: https://www.amazon.com/dp/B09VLHR4JC?tag=hakehardware-20 0:00 - Intro 0:11 - Installing LM Studio 0:53 - Choosing the Right Model for Your GPU 1:51 - Understanding GPU Offload 2:28 - Downloading Your First Model 2:50 - Loading & Configuring a Model 3:52 - First Chat with Local AI 4:35 - Testing Model Quality 5:43 - Adding More Models 7:21 - What's Next Related Videos: - Local AI with a Mini PC is Totally Usable: https://youtu.be/ve4pR7OrEsQ - Host a Local AI with Docker 2024: https://youtu.be/aaURFTEPJVQ - Ollama AMD GPU on Windows — Custom Build: https://youtu.be/AgLL-mmlG9Y - Ollamometer Benchmarks with Strix Halo: https://youtu.be/bnLl1V7VXlU Useful Links: LM Studio: https://lmstudio.ai AI Testing Prompts: https://github.com/copyleftdev/ai-testing-prompts/tree/main #LMStudio #LocalAI #LLM #HakeHardware #AITutorial