Run a policy as an inference endpoint across regions and GPU types, and benchmark it automatically on simulations you build.
We run a policy across cloud regions and GPU types — pinned close to your robots for low latency, with fallback cards for when one type is busy — and benchmark it automatically on simulations you build for the environment it will run in.
Physical AI models behind one key, with no GPUs to spin up. Your robot policy runs as your own deployment — start from a template like π0.5 or bring your own checkpoint — on GPUs close to your robots, with workers that scale with your fleet. The vision-language and decision models it can call share the key.
seq init pi05-droid, then
seq deploy pi05-droid/policy.py — the template becomes a deployment in your own account,
billed to it alone.connect(), every action
goes directly to your own worker over one kept-open connection — the API stays off
the control loop’s hot path.Benchmark a policy automatically on simulations you build for its target environment — every version scored on the conditions it will face, as you iterate. Priced per simulation GPU-hour, at the same GPU rates as inference.
Every model is published the same way — what it reads, what it returns, and what it costs.
vlmvlm
vlmvlm
vlmvlm
vlm
vlmvlmvlm
vlm
decision
Policy templateseq init pi05-droidPolicy templateseq init pi05-libero
Policy templateseq init gr00t-droidPolicy templateseq init cosmos-droidPolicy templateseq init pi05-droid-qwenPolicy templateseq init pi05-robocasaPolicy templateseq init pi05-robotwin
Benchmark
Benchmark
Benchmark
Your own robot policy one seq deploy away, on GPUs close to your robots —
and the vision-language and decision models it can call on the same key. Nothing to provision.