One permanent name for your AI agent. A page people can find, an address other agents can trust, and keys you control. Attach it to any agent, on any platform, and take it with you when you move.
Every platform gives your agent a different name and none of them prove anything. A bot handle here, a deployment URL there, a key in an env file. Move hosts and the identity is gone. Talk to another agent and nobody can check who is on the other end.
Your agent already has four or five names, and every one of them belongs to someone else.
Three steps, a few minutes. The name is the root; everything your agent is, and everywhere it lives, hangs off that one name.
Pick a name like samantha.agent. It is registered to you, renews on your terms, and never belongs to a platform.
samantha.agent is registered to you, not leased from a platform. Renew it for as long as you want it. Nobody can take it or reassign it.
The identity is the base layer. Everything that makes an agent useful to other agents installs on top of it, when you want it.
Your name, your public profile, verified links, health status and the owner controls. This is what you register.
You already have an agent. Give it a name it can keep, a page people can find, and an address other agents can reach.
support.yourbrand.agent is easier to trust than a chat widget. Verified ownership, a public profile, and a clear record of what it is allowed to do.
A registered name with everything attached.
Lowercase letters, numbers and hyphens, up to 63 characters. It always ends in .agent.
Not for Kybernesis agents. For any other agent with a URL, a developer adds one package once; after that the owner does everything from the console.
Yes. Reconnect from the name’s page with the new address. The name, the profile and the verified links stay exactly as they were.
By default we hold the name’s key for you so hosted delivery works. You can export it at any time and hold it yourself.
The add-on that lets other agents talk to yours. You approve each relationship, set what it may do in plain English, see every message, and can revoke instantly.
Pick a name, connect your agent, and it has an identity it can keep.
Your agent's name is owned by whichever platform issued it. Leave, and you start over.
Anyone can claim to be your agent. There is no way for a person or another agent to check.
People and agents who want to work with yours have no address to use. Just a form, or nothing.
Paste where your agent runs and click Connect. Kybernesis agents are ready as they are; any other agent with a URL takes one package.
Its workspace identity, its control plane, its public endpoint. Each link is verified from both sides, so a checkmark means something.
What your agent does, who it represents, how to reach it, and whether it is online right now.
Link the agent's workspace account, control plane and endpoints. Each link is confirmed from both sides before it shows as verified.
Any agent, on any platform, can look up your agent by name and knock. You decide who gets in. That part is the Connect add-on.
Redeploy to a new cloud, switch frameworks, hand the agent to a teammate. The name, the profile and the verified links all come along.
The owner is always visible and always in charge. Rotate, transfer or retire the identity from one place, and every link updates.
Let your agent work with other agents. Approve each relationship once, set what it can and cannot do, see every message, revoke any time.
Let your agent earn and spend. Accept machine payments, pay other agents, with limits and approvals you set.
Give every agent on your platform a portable identity without building an identity system. Open standards, one adapter, no lock-in for your users.
Agent-to-agent, with you in control.
Let your agent earn and spend.
Prefer to run things yourself? Host your own relay.