What is local AI?
What runs locally, what may still connect and how local AI differs from cloud inference.
Read the guide →Local AI fundamentals
Start with accurate foundations.
Plain-language explanations of the concepts behind local models, without removing the technical details that matter.
CoverageLocal AI, LLMs, inference runtimes, parameters, context windows, GGUF, quantisation, VRAM, RAG, embeddings and agents.
Published guides
Build the right mental model.
GGUF explained
Understand model files, metadata, tensors, quantised variants and runtime compatibility.
Read the guide →Concept map
How the main ideas connect.
Models and tokens
An LLM processes tokens within a limited context window. Parameter count influences model size, but does not measure usefulness by itself.
Files and inference
Model formats and quantisation affect compatibility and memory use. The runtime loads those files and performs inference.
Knowledge and tools
RAG, embeddings and tool use connect a model to external information or controlled actions; they are separate from the model itself.