Comprehensive Tool Management with ToolRegistry
ToolRegistry is a protocol-agnostic library designed for managing tools used in function-calling large language models (LLMs). It supports registration of tools from various sources, including native Python functions and OpenAPI specifications, through a unified interface. Key features include concurrent execution with customizable timeouts, a permission system with tag-based policies, and a built-in admin panel for monitoring tool usage and configurations. Its zero-dependency core simplifies installation and usage, requiring only Python 3.10 or higher.
The library facilitates easy integration with multiple LLM providers, such as OpenAI and Anthropic, allowing users to generate tool schemas and execute functions seamlessly. It also supports advanced functionalities like think-augmented calling for enhanced reasoning during tool interactions. With its declarative configuration options and the ability to load tool sources from JSONC/YAML files, ToolRegistry provides a robust framework for developers looking to streamline tool management in AI applications.
