Roundhouse
On the roadmap — register interest

GPU capacity worksheet

The Roundhouse capacity marketplace is on the roadmap, deferred — and no payout rate has been set. So this page won't guess for you. It's a worksheet: you enter a rate, a utilization guess and your electricity price, and it does the multiplication — gross, power cost, and net. Your arithmetic, on your assumptions. Not our projection, and not an offer.

We haven't published a rate, so there's nothing honest to pre-fill. Enter your own assumption; the result is exactly as good as it.

The share of your available hours that would have paid work. That's demand — and we can't predict demand for a marketplace that hasn't opened, so we won't pre-fill this either.

It's on your utility bill. Anyone showing you GPU income without subtracting this is selling you something.

A hardware fact, not an earnings claim. Measure it on your own machine if you can; a mid-range consumer card serving a small open model is often in the tens.

No numbers yet — on purpose

$ —

Fill in a rate, a utilization guess and your electricity price, and this panel shows the arithmetic: estimated tokens per month, gross at your rate, power cost, and net.

We could pre-fill those three fields and show you a dollar figure right now. We won't: the rate isn't set and the marketplace's demand doesn't exist yet, so any number we chose would be our claim wearing your name. When the marketplace has real operating data, we'll publish real earnings distributions here instead.

Where the money comes from

Install one client. Complete jobs. Get paid per token.

Install the client

The same lightweight, pip-installed worker agent that powers corporate on-prem routing. Point it at whatever OpenAI-compatible server your machine already runs — Ollama, vLLM, LM Studio, llama.cpp. Outbound HTTPS only, zero inbound ports.

Complete real AI jobs

The router dispatches inference jobs your models can handle, at the capability tier they qualify for. You choose when the client runs; the router plans around real availability.

Per-token compensation

Every execution is already metered to the token in the routing core. When the marketplace opens, that metering becomes your earnings statement — per token of completed work.

More on the provider side, including the honest state of what's built today: provide capacity.

Like the shape of those numbers?

Tell us what hardware you have — even roughly — and you'll hear the moment provider onboarding opens for the capacity marketplace.

Reserve your spot

No commitment implied by joining — the waitlist is how we plan capacity by region and hardware class.