Magentic: Your Open Source AI Integration Tool

Overview

Magentic is a powerful open-source tool designed to seamlessly integrate Large Language Models (LLMs) into Python code. With its user-friendly decorators, such as @prompt and @chatprompt, developers can create functions that yield structured outputs from LLMs, blending traditional Python coding with advanced AI capabilities.

Features

• Structured Outputs: Utilize Pydantic models and Python types for organized data.
• Streaming Outputs: Get real-time structured outputs and function calls as they are generated.
• LLM-Assisted Retries: Enhance adherence to complex output schemas with intelligent retries.
• Observability: Integrate with OpenTelemetry and Pydantic Logfire for robust observability.
• Type Annotations: Benefit from improved linting and IDE support.
• Multi-Provider Configuration: Easily switch between LLM providers like OpenAI, Anthropic, and Ollama.

Installation

To install Magentic, use:

pip install magentic

or for uv users:

uv add magentic

Configure your OpenAI API key with the environment variable OPENAI_API_KEY.

Usage

With decorators like @prompt, you can swiftly define LLM prompts as Python functions. For example:

from magentic import prompt

@prompt('Add more "dude"ness to: {phrase}')
def dudeify(phrase: str) -> str: ...

dudeify("Hello, how are you?")

Benefits for Users

Magentic simplifies the integration of AI into applications, making it ideal for developers looking to leverage LLM