Orbital 2.0 · Inference at the edge
Deploy any open-weight model to 38 regions with one command. Cold starts under 240 ms, billed per token, observable down to the kernel.
npx orbital deploy llama-4-scout --regions allMove your cursor through the mark