Guides
Learn how to run AI on your own hardware
New to local AI? Begin here.
No prior knowledge assumed. Run your first local model, understand the basic concepts in plain language, and see how local chat grows into an agent — then check the evidence for your own GPU.
Read Start here →Technical guides & deep dives
For readers who already know what they are looking for. You do not need to read Start here first.
How Much VRAM Do You Need to Run an LLM?
A practical guide to VRAM requirements for running Llama, Mistral, Qwen, and other open LLMs locally.
Quantization Explained: Q4 vs Q8 vs Q16
Understand what quantization means, how it affects model quality, and which level to choose.
Single GPU vs Multi-GPU: What You Need to Know
When does it make sense to run multiple GPUs? A practical breakdown of costs, complexity, and performance.
What can I do with my GPU?
When you want evidence rather than explanation, open the primary RigFact journey: what was tested, what happened, and what is still unknown for your hardware.
What can I do with my GPU? →Optional: cloud GPU providers
Local hardware is not the only option. If your GPU does not have the memory you need, or you want to try something before buying hardware, renting a cloud GPU is an alternative. RigFact does not recommend a particular provider, and this is not required to use RigFact.
Vast.ai cloud GPUs →Affiliate Disclosure
Some links on this page are affiliate links. If you purchase through them, we may earn a small commission at no additional cost to you. We only recommend tools we've thoroughly researched and believe add real value.