Run it on our soil, or on yours.
Free to add and pay only for the AI you use, or take the whole engine into your own Azure for a paid license. Both run the same spine and produce the same receipts. The harness and the infrastructure are your choice, and moving between them does not change what the workforce learned.
One engine, deployable inside either boundary.
The same four apps, the same governed engine. Only the boundary around it changes.Add ZeroH from the marketplace and your people are working the same day. The engine runs on our governed fleet, so you never touch infrastructure.
The same engine, dropped whole into your own Azure subscription. Nothing leaves your boundary, and every key belongs to you.
The whole workforce stack. One governed system.
Teams bolt together a stack of separate products to run agents, and still cannot prove what happened. ZeroH ships the whole stack as one governed system, where the proof is a byproduct.
And at the end of it, still no way to prove what happened.
Asked before every pilot.
What do we control when ZeroH runs in our own tenant?
Everything inside the boundary. The whole engine deploys into a managed resource group in your Azure subscription: your Key Vault, your storage, your network and your model provider key. We service the deployment through just-in-time access you approve per request, and no one holds standing administrative rights on your resources.
Do we have to pick a boundary on day one?
No. Start on our fleet in a day with nothing to deploy, then move to your own subscription when you are ready. Same engine, same receipts, same console; nothing to rebuild but the boundary.
Whose AI models does it use?
Yours to choose. Connect your own model provider key or provision Azure OpenAI in your subscription; prompts are masked on your infrastructure before they reach any model, and ZeroH adds no token markup on what the provider charges.