tool-registry

Typed tool catalogue for LLM agents. Register, discover, validate, and export tools to Anthropic/OpenAI format. Zero dependencies.

Install

pip install tool-registry

Quick start

from tool_registry import ToolRegistry

registry = ToolRegistry()

@registry.register(tags=["search", "web"])
def web_search(query: str, max_results: int = 5) -> str:
    """Search the web for information.
    Args:
        query: The search query to execute.
        max_results: Maximum number of results to return.
    """
    ...

# Export to Anthropic format — ready for client.messages.create()
tools = registry.to_anthropic()

# Export to OpenAI format
tools = registry.to_openai()

# Discover by capability
search_tools = registry.find("search")
web_tools = registry.by_tag("web")

# Call a tool
result = registry.call("web_search", query="UK AI regulation 2026")

Auto-inferred from your function signature:

• Parameter types (str, int, float, bool, list, dict)
• Required vs optional (based on defaults)
• Descriptions (from docstring Args section)

---

Linda Oraegbunam | LinkedIn | GitHub