Accountability for autonomous AI.
Evoliv Core builds the identity layer that lets any website verify which AI agent is at its door.
Two projects, one purpose.
Project AIDR
ReleasedA cryptographic identity for every AI agent, verified by websites on their own servers.
Explore AIDRAIAF
Draft 0.1The open standard behind it: who is accountable for an agent, and how strongly that is evidenced.
Read the standardTwo sides of every request.
You run a website
Know which AI agent is calling and who answers for it. Admit it, limit it or refuse it, by your own rules. Free.
Verify agents on your siteYou operate agents
Give each agent an identity websites can check, and that you can revoke in one action.
Explore AIDRHow a request is verified.
The agent signs
With a key that never leaves its device.
Your server checks
Offline. The registry is not in the path.
Revocation is current
A revoked agent is refused within about two minutes.
You decide
Each action sets the assurance it requires.
Open to inspection.
Open-source verifiers
Python, Node.js, Go, WordPress and Cloudflare Workers, under Apache 2.0.
A published standard
AIAF draft 0.1, including what identity cannot prove.
A one-minute quickstart
From nothing to verifying agents on your site.
Evoliv Core is also developing AI for deep medical and chemical research.
Running a website, or running agents?
Verifying is free. Operating agents starts free.