Open Agent Commerce

Get involved

Publish your agent, or get in touch

Open Agent Commerce is built on standards anyone can implement. DNS operators, registrars, identity providers, payment networks, AI platforms, agent builders and security teams are all welcome to adopt and extend it.

Publish your agent

These checklists are deliberately conceptual: the standards move, and the exact steps for your stack will differ. Start here, then follow the standards glossary on the how it works page.

If you run a seller agent

  1. Give your agent a domain (or subdomain) tied to your organisation, and enable DNSSEC for its DNS zone.
  2. Publish a DNS-AID record under that domain, or a catalogue at /.well-known/ard.json, so other agents can discover it.
  3. Obtain a TLS certificate for the agent's HTTPS endpoint, and publish a matching TLSA record so clients can validate it (DANE).
  4. Register your agent in ANS.
  5. Publish an agent card at /.well-known/agent-card.json.
  6. Publish a UCP profile at /.well-known/ucp that declares checkout, the AP2 mandate extension, and the approval services you accept.
  7. Cryptographically sign each quote, and never complete a purchase without a valid, matching mandate.
  8. Maintain a signed ledger of the agent's actions, and periodically commit checkpoints to a public transparency log.

If you run a buyer agent

  1. Register a domain for your agent, or use SPIFFE/SPIRE to establish the identity of the workload it runs on.
  2. Discover and verify sellers before transacting: check DNSSEC signatures, TLSA records and ANS registration, and consider a trust-scoring service.
  3. Keep human approval, and the authority that signs that approval, separate from the buyer agent itself.

Get in touch

Tell us who you are and what you’re interested in. We read every message.