Proof Launches MCP Connector for ChatGPT Users

The first version of Proof's MCP makes its x401 identity protocol available directly to ChatGPT users through the app directory.

Updated on Nov. 18, 2026 in Artificial Intelligence

Bold flat-color editorial illustration of a brass key seated in a metal housing beside a ridged fingerprint pattern, on a navy field.
Proof released an MCP connector letting ChatGPT users approve agent actions with biometric verification through its x401 identity protocol. AI Illustration. Upload story photo >

Live Poll

Would you trust an AI assistant to make payments or sign forms for you?

Proof released the first version of its Proof MCP for ChatGPT, a Model Context Protocol server available as a connector in ChatGPT's app directory. The first flow featured is x401, an open protocol that lets websites and APIs ask an AI agent for evidence of the person behind it.

Why it matters

AI agents can show they have access to a service, but not that a real person approved a specific action. The launch brings a verification layer into the assistant itself, as services from retailers to banks have been blocking agents rather than admitting them.

Proof operates a WebTrust-audited certificate authority and issues verified digital credentials after Kantara-certified NIST IAL2 identity proofing. A hardware-protected key binds a person's verified legal identity to the exact terms of each request, though the extraction gives no adoption metrics for the connector.

The players

Proof

Proof is a company that created the online notarization category and operates a WebTrust-audited certificate authority.

The details

x401 extends HTTP 401, the web's standard unauthorized response, to state what evidence an agent must provide. The person reviews request details and approves with a biometric, and the business verifies the signature itself through the identity wallet.

Timeline

  1. October 15, 2026: x401 launched as an open protocol with production support for dots.

  2. November 18, 2026: Proof released the first version of its Proof MCP for ChatGPT.

The Tech Race

x401 builds on the decades-old HTTP 401 response that signals missing credentials, extending it to tell AI agents exactly what proof of a human's approval is required. The move positions Proof's identity wallet as a verification layer inside the assistant rather than leaving agents to be blocked at the door.

Users of ChatGPT can access Proof's identity flows directly through the app directory connector, with the person approving each request via biometric. Businesses that have been blocking AI agents may now be able to verify a real person's approval before acting.

The takeaway

If an AI agent acts on your behalf, the services it contacts may soon ask for proof that you personally approved the action. Check whether the services you use support identity verification for agent requests.

Further reading

Explore more Artificial Intelligence coverage for context on how AI agents are changing online identity and access.

More information

Read the x401 specification, identity scheme and reference implementation for the protocol details.

Source note: This article includes information reported by The Manila times.

Live Poll

Would you trust an AI assistant to make payments or sign forms for you?