World Adds Agentkit Access as AI Agents Handle Purchases Across 4 Countries

World has expanded access to Agentkit, a development framework that allows individuals to connect AI agents to a verified World ID.

• Key Takeaways:
• • On June 24, World expanded access to Agentkit, enabling AI agents to link directly to verified World IDs.
• • A trial of 500 hats proved that platforms using Agentkit can stop automated retail bot abuse.
• • World aims to build a global “agent economy” by integrating Agentkit across digital services

Tackling the Bot Problem

Sam Altman-backed World announced June 24 it is expanding access to Agentkit, a development framework that allows individuals to link artificial intelligence (AI) agents to a verified World ID.

The framework’s latest expansion builds on two major foundational announcements rolled out by the company earlier this year. World initially unveiled the technology in March, launching the first Agentkit beta. Developed in collaboration with Coinbase and Cloudflare, the initial release integrated World’s biometric identity layer with the x402 protocol, a payment rail designed to allow human-backed AI agents to securely execute stablecoin micropayments.

One month later, the company said it had embedded the infrastructure into its full-stack World ID 4.0 protocol upgrade. That unveiling expanded the agent ecosystem by establishing official developer integrations with core internet infrastructure and enterprise platforms, including Vercel, Okta, Box and Browserbase.

According to a media statement, the latest rollout addresses an emerging security challenge for online businesses. As AI agents grow capable of autonomously handling complex tasks, companies increasingly struggle to distinguish between a legitimate agent representing a single consumer and malicious bot networks. By anchoring AI agents to World ID, the framework aims to provide an identity-based trust layer.

It allows users to delegate complex digital workflows to automated tools while giving businesses a mechanism to verify that each agent maps back to a unique and verified individual. Setting up the system requires a verified World ID, the World app and a compatible AI agent platform, with current support including Claude Code, Codex, Cursor, Hermes, and Openclaw.

Users link their digital proof of humanness through World’s Toolrouter interface to generate an API key, a process the company states takes only a few minutes. To test the integration at scale, the company recently conducted a pilot program featuring a limited-edition release of 500 “Human in the Loop” hats, which were restricted to verified World ID holders.

During the trial, users’ AI agents independently tracked the product launch, verified their owners’ eligibility, navigated the digital storefront, and finalized the transactions. And because the storefront was integrated with Agentkit, the platform successfully enforced a strict one-item-per-person limit across international borders, including orders completed in the United States, Germany, Japan, and the United Kingdom.

Company officials said the trial demonstrates how e-commerce platforms can safely open their infrastructure to automated buyers without risking inventory depletion by automated resale bots. As Agentkit integration expands to broader digital services, the company aims to build an accountable “agent economy,” ensuring that autonomous software remains bound to and controlled by the humans they represent.