Tools for Humanity, the startup behind Sam Altman's World project, has released AgentKit in beta — a developer toolkit that lets AI agents carry cryptographic proof they are acting on behalf of a real, verified human.

How it works

AgentKit links a user's World ID (derived from an iris scan via World's Orb device) to the x402 v2 protocol, the open standard for machine-to-machine payments developed by Coinbase and Cloudflare. Once a user registers their AI agent with their World ID, the agent can attach that proof to any interaction with an x402-compatible site.

The result: a platform can verify that a purchasing agent represents a unique human without learning anything else about that person. "It's not necessary for the identity part to contain information about the individual themselves — we're purely anonymous in the proof-of-human protocol," DC Builder, a World Foundation research engineer, told Decrypt.

Why it matters

The problem is concrete. AI agents are increasingly used to book reservations, buy concert tickets, and access APIs. Without identity signals, a single bad actor could deploy thousands of agents to scalp tickets or spam services. Federal courts have already stepped in — a judge recently blocked Perplexity's Comet browser from autonomously purchasing on Amazon.

AgentKit is designed as a complementary layer to x402, not a replacement. Any site already accepting x402 payments can add proof-of-human verification alongside micropayments.

World says its network currently includes nearly 18 million verified individuals across more than 160 countries. WLD tokens are not required to use AgentKit.