The World project has introduced its “most significant update” to the World ID protocol, positioning the system as a proof of human infrastructure aimed at consumers, businesses, and AI agents.
New Architecture
The upgrade is based on a shift to an account-based architecture: rotation and support for multiple keys, access recovery, and session management make the protocol suitable for corporate use.
One-time nullifiers are designed to enhance anonymity guarantees, while a new open-source SDK allows any application to act as a World ID authenticator.
Simultaneously, the project launched the World ID app, specifically designed for managing and using online identity verification. Initially, the product is available as a public beta version.
Project Partners
According to World, several consumer platforms have continued to integrate proof of human infrastructure to enhance trust:
- Tinder — expanded World ID implementation to the US. Verified users receive a badge on their profile and five free Boosts. Previously, the partnership was active in other markets;
- Zoom — became the first communication platform to integrate Deep Face directly into calls. It provides three-way identity matching: Orb verification, real-time Face Auth selfie, and current video frame;
- Docusign — document signers confirm through World ID that they are a living person, not a bot;
- Vercel — integrates human in the loop for AI agent creators and workflows using the company’s SDK. Any step in the action chain can require human verification—with full auditing;
- Okta — plans to launch the Human Principal product, allowing API developers to verify if a real person is behind an agent and apply policies based on this. World ID is a key integration partner;
- Match Group — implements Razer ID technology based on World ID as a standard for games;
- Mythical Games — uses user authentication for the gaming economy.
Focus on AI Agents
World emphasized that the update prepares the infrastructure for a new class of users—AI agents. The protocol will allow verification that each assistant is backed by a unique human, which becomes critical as autonomous systems spread.
Back in March, World introduced the AgentKit product for identity verification in the agent internet.