I build gpu.nz, a per-minute GPU cloud. Writing about VRAM, local LLMs and running ComfyUI, Ollama and vLLM on rented GPUs.
About
gpu.nz is a GPU cloud billed by the minute: RTX 3060 to H100/H200/B200 with ready ComfyUI, Ollama, vLLM and Jupyter templates. I write about how much VRAM open models really need and how to run them on rented GPUs. Free VRAM calculator: https://gpu.nz/en/tools/vram