Welcome to Genie Tooling

Genie Tooling is a hyper-pluggable Python middleware designed to empower developers in building sophisticated Agentic AI and LLM-powered applications. It provides a modular, async-first framework that emphasizes clear interfaces and interchangeable components, allowing for rapid iteration and customization of agent capabilities.

Key Features

• Simplified Development with the Genie Facade: The primary entry point for most applications, offering a high-level API for common agentic tasks like LLM interaction, Retrieval Augmented Generation (RAG), tool execution, and natural language command processing.
• Hyper-Pluggable Architecture: Almost every piece of functionality—from LLM providers and tools to data loaders and caching mechanisms—is a plugin. This allows you to easily swap, extend, or create custom components using a simple Bootstrap Plugin system.
• Explicit Tool Enablement: For production safety, tools are only active if explicitly listed in the configuration. The auto_enable_registered_tools flag (True by default) allows for rapid development by automatically enabling tools registered via the @tool decorator.
• @tool Decorator: Effortlessly convert your existing Python functions into Genie-compatible tools with automatic metadata and schema generation.
• Zero-Effort Observability: The framework is deeply instrumented. Enabling a tracer (e.g., console_tracer or otel_tracer) provides detailed, correlated traces for all internal operations through a pluggable LogAdapterPlugin system.

Quick Start

Get up and running quickly with the Genie facade:

import asyncio
import logging
from genie_tooling.config.models import MiddlewareConfig
from genie_tooling.config.features import FeatureSettings
from genie_tooling.genie import Genie

async def main():
    logging.basicConfig(level=logging.INFO)

    app_config = MiddlewareConfig(
        features=FeatureSettings(
            llm="ollama", # Default: ollama, ensure it's running
            llm_ollama_model_name="mistral:latest",
            command_processor="llm_assisted",
            tool_lookup="hybrid"
        ),
        # Explicitly enable the built-in calculator tool.
        # For production, set auto_enable_registered_tools=False
        # and list all tools here.
        tool_configurations={
            "calculator_tool": {}
        }
    )
    genie = await Genie.create(config=app_config)
    print("Genie initialized!")

    # LLM Chat
    chat_response = await genie.llm.chat([{"role": "user", "content": "Tell me about Genie Tooling."}])
    print(f"LLM: {chat_response['message']['content']}")

    # Command Execution
    cmd_result = await genie.run_command("What is 5 times 12?")
    print(f"Command Result: {cmd_result.get('tool_result')}")

    await genie.close()
    print("Genie torn down.")

if __name__ == "__main__":
    asyncio.run(main())

Dive Deeper

• User Guide: Learn how to install, configure, and use Genie Tooling for your projects.
  • Installation
  • Configuration & Simplified Configuration
  • Using LLM Providers
  • Using Tools
  • Using RAG
  • Using Command Processors
  • Logging & Observability
  • Distributed Tasks
• Developer Guide: Understand the plugin architecture and learn how to create your own custom plugins and tools.
  • Plugin Architecture
  • Extending Genie Tooling (Creating custom extensions)
  • Creating Tool Plugins
  • Creating RAG Plugins
  • Creating Other Plugins
• API Reference: Detailed reference for all public modules and classes.
• Tutorials & Examples: Step-by-step guides and practical examples to get you started.

We encourage you to explore the documentation and examples to unlock the full potential of Genie Tooling!