OpenVerb

OpenVerb is an open standard designed to bring clarity and structure to AI actions within software systems. It addresses the challenge of hidden or scattered AI capabilities by providing a unified system where every possible action—referred to as a "verb"—is explicitly named, defined, and made visible. This approach allows developers to understand, control, and orchestrate the actions their AI agents can perform, moving away from opaque, buried instructions toward a transparent and auditable framework.

The platform introduces a system in which each action has a clear name, defined inputs, outputs, and effects. OpenVerb enables developers to register, execute, and audit these verbs, ensuring that nothing is hidden within prompts or fragmented across codebases. The process includes loading a verb library using a JSON schema to define available actions, allowing the AI to propose structured actions based on this vocabulary. Before any action is executed, OpenVerb Core validates the action by checking its existence in the library, ensuring the payload matches the schema, verifying policy requirements, and sanitizing inputs. Once validated, the action is routed to the backend for execution, and an execution receipt is generated for full auditability.

OpenVerb is particularly relevant for developers building AI-powered applications who require transparency, control, and traceability in their systems. Its definitive implementation for Python, called OpenverbFellowship, can be installed via pip, making it accessible for Python-based AI agents. The system is designed to work alongside existing tools, revealing and organizing capabilities rather than replacing them.

By making AI actions explicit and auditable, OpenVerb aims to transform the development and management of intelligent agents, fostering a more understandable and manageable environment for scalable AI systems.