World Launches AgentKit That Leverages x402 Protocol to Power Human-Backed AI Agents

World, the network responsible for the proof of humanity solution World ID, has just unveiled a major new product. AgentKit is targeted at developers, but web users will be the primary beneficiaries of the technology once it’s integrated into apps focused on the agentic economy. The tooling makes clever use of the x402 protocol created by Coinbase and Cloudflare to connect real people with online agents.

WorldKit marks World’s first serious foray into agent frameworks, despite the World ID system having been originally engineered to allow humans to prove their identity in the digital world. WorldKit brings this capability to agents by proving that each agent is controlled by a unique human. It’s an attempt to offer the best of both worlds: freedom to deploy agents to undertake tasks on your behalf, bolstered by the ability to prove that you’re the one controlling them.

World ID on a Web Awash With Agents

If you’re the operator of a web platform or mobile app, it can be hard to tell who your customers are. Is the entity sending support requests and using your API a human or are they the product of a bot farm, running thousands of automated accounts? That’s one of the primary problems AgentKit has been designed to solve. The idea is to foster a web of human-backed agents, each associated with a unique human.

The technology uses cryptographic proofs, like those already built into World ID, that allow an individual to prove they are real without revealing valuable personal data. The x402 protocol that AgentKit is built on is centered around payments for agents, but with AgentKit plugged in, it’s now possible to determine who is behind the agent running the task in question.

According to x402 Founder Erik Reppel, who is also Head of Engineering at Coinbase Developer Platform, integrating World ID with the x402 protocol gives developers “a complete trust stack: a way for agents to pay for what they need and a way for platforms to verify there is a real human behind the wallet. This is a massive step toward a web where agents aren’t just seen as automated traffic, but as legitimate economic participants.”

Less Bots, More People

From a technical perspective, AgentKit functions as an extension of the x402 protocol, embedding identity verification directly into the request-response cycle between agents and online services. When an agent interacts with a compatible website or API, it can present both a micropayment and proof of a unique human.

The latter is generated via World ID using privacy-preserving cryptography, allowing the receiving platform to confirm that the agent is backed by a distinct individual without accessing any personal data. This dual-authentication model, in which payment governs access while identity governs legitimacy, enables platforms to differentiate between high-volume bot traffic and genuine user interactions.

Use cases envisioned for this technology include allowing restaurants to accept table bookings made by agents with confidence that a real human will show up for dinner. Likewise, when it comes to purchasing concert tickets or flights. And of course, the tech should prove its worth in the onchain economy too, allowing users to interact with crypto platforms that require verification without needing to share sensitive data.

Initially available in beta, AgentKit looks poised to drive greater uptake of AI agents, not least from the World’s existing 18 million verified users, who’ll soon be able to automate tasks while proving that the request is genuine and the operator is very real.