OpenTerms

OpenTerms establishes an open standard for defining and checking machine-readable permission rules that govern what actions AI agents may perform on a given service. It addresses the problem of AI agents making incorrect assumptions about their permissions, which can lead to unauthorized actions, workflow failures, or unnecessary interruptions. By providing a clear and deterministic way for agents to verify permissions before acting, OpenTerms aims to prevent issues such as agents being blocked for forbidden actions, mistakenly performing actions that should be denied, or hesitating when they actually have permission.

well-known/ path, specifying what actions are allowed, denied, or require escalation for agents. Agents can then use the OpenTerms SDK to check permissions programmatically with a single call, receiving responses such as allow, deny, or escalate. This process is entirely deterministic, relying on a direct lookup rather than AI inference, ensuring consistent and predictable results. com," is permitted. The standard is designed to be vendor-neutral and built on published cryptographic standards, including RFC 8785, RFC 7517, Ed25519, and JCS, allowing anyone to verify compliance without proprietary lock-in.

OpenTerms is positioned as an open standard, and its SDK is free to use, with no lock-in for developers or service providers. The platform currently indexes hundreds of services, making it broadly applicable for those looking to implement permission checks for AI agents. Additional components, such as signed receipts for action verification and delegated-authority commitments, are mentioned as part of a broader evidence chain, though some are in limited release. OpenTerms is suited for developers, engineers, and organizations seeking to manage and automate permission compliance for AI agents in a transparent and standardized manner.