Local AI

Run AI on
your own hardware.

Learn local AI from first principles: hardware and VRAM, quantization and GGUF, then move into a practical runtime with Ollama.

Local AI runtime pipeline from model file through runtime and memory to token generation
Use the cluster as a path from model fundamentals to a measured local configuration.

Start here

1 — What Is Local AI?2 — Local LLM Hardware Requirements3 — LLM Model Size Explained4 — How Much VRAM Do You Need for Local LLMs?5 — Choosing a Local LLM Quantization6 — What Is GGUF?7 — Context Length in Local LLMs8 — What Is Ollama?9 — CPU vs GPU for Local LLMs10 — How to Run AI Models Locally11 — Model Offloading for Local LLMs12 — Choose a Local AI Model for Your Hardware13 — Managing Local Models with Ollama14 — GPU Offloading in Ollama: What It Actually Means15 — How to Check Model Memory Usage in Ollama16 —
Learning path

Local AI foundations

Navigation is generated from article metadata and stays synchronized as the cluster grows.

Run

Model → runtime → tokens

See how the complete local inference configuration fits together.

Fit

When VRAM is not enough

Understand mixed CPU/GPU placement using our measured Qwen example.

Choose

Hardware-aware selection

Turn memory, context and workload constraints into a repeatable workflow.

Interactive planning

Will the model fit your hardware?

Estimate GPU fit, partial offload or RAM-heavy execution before you load the model.