/waitlist

Hosted compute comes later.

Hosted management does not mean hosted workers are available. Hosted compute means the agents will run on my AWS instead of yours: same product, more expensive, and EC2 stops being your problem.

Two different things get called hosted.

Planned first

Hosted management

RepoStaff will host the Management UI and the integrations on every plan, including self-run. Your agents will still run on compute you choose: your Linux/KVM hardware first, then your AWS.

Later. Waitlisted. No date

RepoStaff-hosted compute

Marshall and the ephemeral agents will run on RepoStaff's AWS. Gated, and only after the paid self-run launch and its readiness checks. This is the page you're on.

The plan, if it opens.

Planned, untested hypotheses

$699 per org, per month
100 VM-hours included, then $0.60 per VM-hour up to a required per-org overage cap
Card on file, and no free trial on hosted compute
Your own Anthropic or OpenAI API keys, on a separate invoice from your provider
That cap covers your RepoStaff bill only. It doesn't limit your model provider's invoice