About · Sheet 02
We build inference infrastructure the way we wish we'd had it.
Seattle Compute Company is based in Seattle, Washington — close to the water, occasional sun, and more coffee than is strictly necessary.
Origin
Why dedicated inference?
Shared inference APIs are great right up to the moment they aren’t. You hit a rate limit you didn’t know existed. A noisy neighbor tanks your p99. Costs swing 5× between a quiet week and a big launch.
Dedicated inference fixes all of that — at the cost of building and babysitting GPU infrastructure. Seattle Compute handles the second half of that trade so you can have the first half.
We’re a small team that cares about sane defaults, fair pricing, and letting the open-source model ecosystem win.
Values · Sheet 03
No. 01
Transparency over margin
You see the quote before you deploy, and the number on your bill matches the number in the dashboard. We’d rather explain our margin than hide it.
No. 02
Infra that gets out of the way
The right GPU is a solved problem given a model, a precision, and a priority. We solve it for you so your team can focus on the product.
No. 03
Model-agnostic by default
Any open-source model from HuggingFace or Ollama. No lock-in, no “foundation model” upcharge. The ecosystem is moving fast — we think your infra should too.
Contact
Say hello.
Questions about your workload, pricing, or a specific model? We read every email.