Straight answers about running your own AI.
Every number here is a measurement, a stated formula, a price read from the provider with its date, or yours.
Where to run
Every price with the date we read it, and what the hourly rate leaves out.
Rent a GPU: cloud GPU prices compared On-demand prices from 13 providers for every card from the RTX 4090 to the B200, the card your model needs, the price per token, and what the hourly rate leaves out. 15 min read ComfyUI in the cloud, compared Comfy Cloud, RunComfy, ThinkDiffusion, MimicPC, Floyo, RunPod and Vast.ai: price per hour by card, what idle time costs, nodes, files and privacy. 12 min read ComfyUI API: your options compared Comfy API, RunComfy, ViewComfy, fal, Replicate, RunPod Serverless and Modal: what a GPU hour and one job cost, custom nodes, cold starts, agent spend limits and privacy. 12 min read Ollama Cloud: pricing, models and privacy What a month of credits buys on each cloud model, which ones fit a machine at home, and what Ollama says about your prompts. 11 min read
Run it yourself
What fits your card, which machine to get, and what it costs to run.
What is a private AI server? What "private" means at each layer, and the three ways to get one, honestly compared. 9 min read Self-hosted LLM: the practical guide How VRAM is really used, the runtimes that matter, quantisation without the mysticism, and a calculator. 11 min read Local AI server: build or rent? The three build tiers, the running costs nobody lists, and a break-even calculator with your own numbers. 9 min read OpenClaw VRAM requirements The context OpenClaw asks for, what it costs in memory, and the best tool-calling model for 12, 16 and 24 GB. 7 min read Personal AI computers compared Ghost Core, DGX Spark, Strix Halo, Mac Studio and RTX PCs: the memory each GPU can use, its speed, and the models each one runs, agents included. 8 min read DGX Spark vs Strix Halo vs Mac Studio The three big-memory boxes for local LLMs: memory the GPU can use, bandwidth, software, and speed on the same models. 7 min read Which models fit on 16 to 96 GB Estimated VRAM for popular open models at Q4 and Q8, 8k and 32k context, with the smallest card that fits. 6 min read
Free tools
No sign-up, and they run in your browser.
ComfyUI workflow checker Drop a workflow: every model file it loads, the VRAM it needs, what to swap and the custom nodes. Can I run it? Your card, any model: fits or not, how fast, and what to change when it does not. Build vs rent calculator Your hours, your electricity, your plan. Finds the break-even month.
Every model, card and place to run
Where to run, compared Your own GPU, Comfy Cloud, a cloud GPU by the hour or a per-clip API: what each costs, and when each one wins. Best GPU for AI Every card against the same open models: the pick at each memory size, what fits at Q4 and how fast each answers. 39 GPUs LLM VRAM requirements Every open model measured the same way: memory at Q4 and Q8, what each 1,000 tokens of context costs, and the smallest GPU that runs each one. 34 models Image and video models How much VRAM open image and video models need in ComfyUI, worked out from the exact files. 11 models
Nobody pays to be listed or ranked: how we stay independent.