What is local AI?
A practical definition, including what stays local and what may still use a network.
Read the guide →Loco Learn is a structured knowledge base for researching models, hardware, runtimes, privacy, benchmarking and practical deployment.
Master artificial intelligence from first principles, understand neural networks and LLMs, and learn how to run models privately on your computer.
Each hub connects foundational explanations, practical guides and evidence-focused research.
Comprehensive structured courses from AI 101 to local LLM mastery.
01LLMs, inference, GGUF, quantisation, context, RAG, embeddings and agents.
02Choose by task, hardware fit, runtime compatibility and licence.
03Understand methodology, performance measurements and evidence labels.
04RAM, VRAM, CPUs, GPUs, laptops, workstations and servers.
05Model formats, hardware acceleration, interfaces and integration choices.
06Private AI, on-premise deployment, governance, cost and data considerations.
07POPIA considerations, local costs, infrastructure and business adoption.
A practical definition, including what stays local and what may still use a network.
Read the guide →What GGUF stores, how it relates to runtimes and what a filename cannot tell you.
Read the guide →Estimate model memory and leave room for context, runtime and the operating system.
Read the guide →A grounded look at opportunity, infrastructure, privacy and organisational responsibility.
Read the guide →Loco intends to publish at scale without manufacturing expertise. Technical specifications are checked against current primary sources. Original tests describe their hardware and methodology. Third-party results stay labelled as third-party.