The default: hosted and isolated
By default Agent Hive hosts your colony as an isolated instance with its own database and volume. This already gives you the no-noisy-neighbor guarantee and keeps your business data out of any shared store.
The self-host path
The engines under the platform are open and permissively licensed, which makes a self-host path possible for teams that require the keys and the data to stay entirely within their own infrastructure. The open-source page lists the engines, their repositories, and their licenses.
Self-hosting is most relevant to regulated and public-sector teams. The exact deployment shape and support model are being finalized. Talk to us if this is a requirement for your business.
Data residency questions, answered honestly
For most teams, the hosted-and-isolated default already answers the data-residency question: your business data lives in your colony's own database, not in a shared store, and the control plane holds only account and billing identity. When a reviewer asks where customer data goes, the honest answer is that it stays in the colony.
Where teams need more, the open foundation is the lever. Because the engines are open source, the path to running the whole stack inside your own network exists rather than being a closed door. We would rather tell you plainly what is available today and what is still being scoped than overpromise a self-host story that is not finished.
When to choose self-hosting
Self-hosting is the right call when a contract or a regulator requires that data and keys never leave infrastructure you control, or when an internal policy forbids a managed model key. It is the wrong call if you are reaching for it out of general caution: the hosted isolation model already gives you a strong, defensible separation, and self-hosting adds operational burden you should only take on when you actually need it.
This article is part of the launch docs set; boundaries and depth are still being reviewed with engineering and will keep sharpening.
