Skip to content
Deploy

AI and GPU server guides

Guides to running AI models on a server you control: how to self-host an open-weight language model, how to choose between Ollama and vLLM, how much VRAM a model needs, and how to keep a model endpoint private. Commands come from each project's own documentation, which every guide links, and model sizes are worked out from arithmetic, not from benchmarks.

Latest posts

The guides here run on a GPU server with NVIDIA RTX PRO 6000 Blackwell GPUs.

See GPU server plans