Tell us about your use case
Request access, or tell us what your agents do, or should be doing. We’ll get back to you within one business day.
What happens next
- We email you to set up a 30 minute call.
- We go through a real case together: which agents, which actions, who decides.
- We set up a room with your first policies, ready to try.
Loading the form…
Questions people ask
What counts as a decision?
Any action an agent proposes that gets decided: approved or rejected by a person, approved or blocked by the arbiter, or run on autopilot. Chat messages don’t count.
Do I have to rewrite my agents?
No. Add one HTTP call at the point where the agent is about to act. In n8n that’s an HTTP Request node; in Python or Node it’s a few lines.
What happens when I run out of decisions for the day?
New proposals are refused with a clear error and an event in the room, so the agent knows to wait. The count resets the next day.
Who pays for the AI?
You can connect your own Claude, OpenAI, Azure OpenAI, Gemini, Mistral or any OpenAI-compatible account: requests go from otorize straight to your provider and you pay them, with no daily cap. The key stays encrypted and is never shown again.
How do I get an account?
On request: fill in the form on the Contact page. We set up every account by hand, to keep out abuse and fake accounts, and email you as soon as you can log in.
Where is my data?
On Free and Team, in a shared database in the EU, with every organisation kept apart at the database level. Enterprise gets a dedicated database. Either way, messages, action parameters and emails are encrypted with a separate key for each organisation, so anyone reading the database sees nothing usable.
Can I run it on my own servers?
Yes, on the Enterprise plan. We ship it as a container and help you set it up.