VercatusVisa trusted agent protocol
The agent signs each request to the merchant with an RFC 9421 HTTP Message Signature, aligned with Web Bot Auth, and the merchant checks it against Visa's published public keys; the signature's tag says whether the agent is browsing or paying.
How does Visa's Trusted Agent Protocol tell a merchant that an AI agent is genuine?
Visa's Trusted Agent Protocol (TAP, 14 October 2025, built with Cloudflare and published on the Visa Developer Center and GitHub) has a trusted agent sign each request to a merchant with RFC 9421 HTTP Message Signatures aligned with Web Bot Auth. The Signature-Input tag states whether the agent is browsing (agent-browser-auth) or paying (agent-payer-auth); public keys come from https://mcp.visa.com/.well-known/jwks (Ed25519, PS256 or ES256); created and expires may span at most 8 minutes and a nonce must not be reused. Optional signed objects carry consumer recognition (agenticConsumer, with an ID token holding obfuscated email or phone) and payment data (agenticPaymentContainer). Akamai announced edge verification of TAP with Visa on 17 December 2025.
Sources
- 1
- 2Visa · 2025Visa introduces Trusted Agent Protocol: an ecosystem-led framework for AI commerceusa.visa.com
- 3
AGrade A: official documentation or a standard. A grade describes the provenance of the sources above; it is not a probability that a change will work. The effect basis says how the size of the effect is known: measured by a study, taken from an industry report, only documented without a size, or not quantified.
Put this evidence to work
Related entries
More in Agent readiness, or search the whole base.
All Agent readiness entries