Skip to content

ToolRegistry Ecosystem

The ToolRegistry ecosystem consists of three complementary packages that work together to provide a complete tool management solution for LLM applications.

Overview

Package Description PyPI Docs
toolregistry Core library for protocol-agnostic tool management PyPI Docs
toolregistry-server Server adapters (OpenAPI & MCP) for exposing tools PyPI Docs
toolregistry-hub Curated collection of ready-to-use utility tools PyPI Docs

Dependency Diagram

toolregistry          ← Core: tool registration, schema generation, execution
toolregistry-server   ← Server: OpenAPI & MCP protocol adapters
toolregistry-hub      ← Hub: ready-to-use tool implementations

Package Details

toolregistry (Core)

The foundation of the ecosystem. Provides:

  • Protocol-agnostic tool registration and management
  • OpenAI-compatible function calling schema generation
  • Concurrent tool execution with multiple modes
  • Integration adapters for MCP, OpenAPI, LangChain, and class-based tools

toolregistry-server (Server)

Built on top of toolregistry. Provides:

  • OpenAPI REST adapter for exposing tools as HTTP endpoints
  • MCP (Model Context Protocol) adapter for AI-native tool serving
  • Route table for organizing and managing tool endpoints
  • Authentication and configuration support
  • CLI for quick server deployment

toolregistry-hub (Hub)

A curated set of utility tools, deployable via toolregistry-server. Provides:

  • Calculator, unit converter, date/time tools
  • File operations and file system tools
  • Multi-engine web search (Brave, Tavily, SearXNG, etc.)
  • Web content fetching, think tool, todo list
  • Docker image for one-command deployment

This Project

You are currently viewing the documentation for toolregistry-server — define custom tools and serve them via OpenAPI or MCP interfaces.