World Network AgentKit links AI agents to human IDs

World Network has expanded access to its AgentKit framework, a system designed to connect autonomous AI agents to verified human identities. The rollout aims to address growing concerns about distinguishing legitimate software acting for individuals from large-scale automated bot networks as AI agents take on more online tasks, including making payments and purchases.

The framework links supported AI agents to a user's World ID, a digital identity verification system. This allows websites and applications to confirm that an agent is operating on behalf of a unique, verified human. According to the company, this identity layer is intended to help businesses enforce user-level rules, such as purchase limits, and reduce abuse from automated accounts.

To use AgentKit, an individual needs a verified World ID, a World App account, and a compatible AI agent. Supported agents include Claude Code, Codex, Cursor, Hermes, and OpenClaw. Users can generate credentials and authorize their agents through a ToolRouter interface, enabling the software to interact with supported services while maintaining identity verification tied to the user's account.

World recently demonstrated the technology...