Local & cloud LLM hosting

Find a model for your hardware

Start with the hardware you have, or compare a cloud server. Find reviewed setup guides with exact model versions, memory assumptions and costs.

No account requiredSource-backed configurationsUnknowns stay visible

What will you run it on?

Choose your computer, then tell us what is available. We’ll compare reviewed setups and show exactly what still needs checking.

Apple computerMacBook, Mac mini or Mac StudioPC with a graphics cardChoose your NVIDIA or other GPUCPU only or unsureStart with what you know

Operating system

Must match the guide’s documented operating system. Windows and WSL2 are checked separately.

Available GPU / unified memory (GiB)

Free memory on one graphics card, after other apps. Do not add multiple cards together.

Available system RAM (GiB)

Memory left for the model and runtime after other programs. Leave blank if unknown.

Free storage available (GiB)

Space you can use now for downloads and the runtime, not your disk’s total size.

Manual setup and advanced requirements

Hardware configuration

For other devices or cloud servers. This replaces the device preset; the guide’s complete setup still applies.

Memory and disk units

Use the unit shown by your system. GB and GiB are different.

What will you use it for?

Context length (tokens, optional)

Optional: how much text you need at once. Leave blank to use each guide’s documented context and memory requirements.

We don’t inspect your device. Blank amounts stay unknown. Change GB / GiB units under advanced requirements. A memory fit does not establish speed, model quality or runtime support.

Choose your requirements and select Find configurations. Missing hardware evidence will be shown as unknown.

Compare hosting configurations

Compare two exact setups, including different configurations of the same model. Hardware fit, measured speed and costs are separate decisions.

First configuration

Second configuration

Add your cost assumptions (optional)

Use hours for the same scenario period. Cloud billing includes idle time; active hours alone do not set the bill. Local electricity uses whole-system attributable power, powered hours and your tariff. Missing inputs remain unknown.

Cloud billed hours

Active workload hours

Administration hours

Local powered hours

Attributable power (W)

Electricity rate / kWh (guide currency)

Stored data (GB)

Outgoing data (GB)

Choose two different configurations to see their tradeoffs.

Understand the evidence

Documented support means a setup has sources. Estimated fit shows its assumptions. Measurements apply only to the exact hardware, runtime and workload tested.

Keep costs separate

Hardware and setup are one-time costs. Cloud uptime, electricity, storage and transfer are recurring. Missing prices are unknown, never free.

Choose your next step

Open a configuration’s guide for prerequisites, commands, verification and shutdown. Compare model quality and API pricing in the existing catalog.