World has expanded access to AgentKit, a framework that has enabled verified users to connect AI agents with their digital identities and demonstrate that those agents represent real people rather than automated botnets.
Summary
- World has expanded AgentKit, allowing AI agents to operate on behalf of human users verified through World ID.
- The framework supports AI tools such as Claude Code, Codex, Cursor, Hermes, and OpenClaw.
- A recent sale of 500 limited-edition caps demonstrated how verified AI agents can complete purchases while also enforcing single-person limits.
According to World, the launch comes as AI agents take on a growing number of online tasks, increasing demand for systems that can distinguish between software acting for a specific individual and large-scale automated operations.
The project, backed by OpenAI CEO Sam Altman, said AgentKit allows users to delegate actions to AI tools while maintaining identity verification and user controls.
The launch follows a period of increased attention to the project following major US crypto trading platform Robinhood. list the global token, bringing additional visibility to the ecosystem.
Verified identities allow AI agents to act on users’ behalf
Details published by World show that AgentKit links supported AI agents to a verified global ID, allowing websites and apps to confirm that an agent is acting on behalf of a unique human user. The company said the system is designed to help companies enforce rules at the user level and reduce abuse associated with automated accounts.
To access the framework, World stated that users need a verified World ID, a World app account, and a compatible AI agent. Currently supported options include Claude Code, Codex, Cursor, Hermes, and OpenClaw.
Through World’s ToolRouter interface, users can create credentials and authorize their agents to interact with supported services. According to the company, this process allows people to assign tasks to artificial intelligence systems without giving up identity verification linked to their accounts.
Instead of relying solely on account credentials, the framework adds evidence that an agent represents a verified person, which World says can help online services distinguish legitimate activity from coordinated bot behavior.
Demo shows AI agents completing purchases
To showcase the technology, World recently hosted a limited-edition sale of 500 “Human in the Loop” hats. According to the company, AI agents took care of the entire purchasing process for participating users.
World said agents discovered the product launch, verified eligibility requirements, navigated the online store and completed transactions without direct user involvement during the purchase flow.
Identity checks remained active throughout the event. According to World, World ID verification ensured that purchase limits applied to one person per item, preventing users from bypassing restrictions through multiple automated accounts.
The company reported that the 500 hats were claimed by verified users located in countries such as the United States, Germany, Japan and the United Kingdom. World said the event demonstrated how companies can enable AI agents to perform online actions while maintaining controls aimed at reducing bot-driven abuse and automated farming activity.
As AI-powered software takes on more responsibilities on digital platforms, World said AgentKit provides a way to connect those agents with verified human identities, allowing organizations to verify who is ultimately behind automated actions carried out online.
