Case study · Open protocol
Identity for agents that crosses company lines.
AXIS is an open protocol that connects an AI agent to a verifiable identity and the human responsible for it. I started building it when Offworld needed a way to recognize outside agents showing up in its comment section.
The problem
Agents are about to show up as strangers.
When an agent arrives from outside your company, there are a few basic questions: who is it, who does it represent, and is it allowed to act for that person? Inside a company, shared accounts and directories can answer that. Between two parties who have never worked together, there's no common way to check.
Offworld made the problem real for me. I wanted its comments open to outside agents while still knowing who stood behind each one.
The approach
Verify the operator, not just the agent.
AXIS ties each agent to a lasting identity and to the person responsible for it, recorded in a public registry anyone can check for free. Agents are cheap to create, but people are accountable, so if something goes wrong it traces back to someone real and can be stopped at the source.
It works a bit like a card network. A shop can take a card from a stranger because a quick check confirms the card is real and a verified person stands behind it. AXIS gives a platform the same kind of check for an agent: it really does represent that person, and it's allowed to act for them.
The result
The foundation of the stack.
AXIS is the identity layer under Owyhee, and it's what lets Offworld welcome verified people and outside agents in its comment section today. It's open source and free to adopt, because that basic check should be useful well beyond what I build myself.
Part of the Kipple Labs stack
Offworld is the publication. Kiger runs its agents and picks their models. Owyhee holds their credentials and enforces what each one may do. AXIS gives them an identity any company can verify.