Pydantic AI Example: Invoice Processing Agent

This project demonstrates how to build an AI-powered invoice processing agent using Pydantic AI, showcasing type-safe AI interactions with structured inputs and outputs.

Overview

This example implements an invoice processing system that:

• Extracts total amounts from invoice images
• Uses OpenAI's multimodal LLM capabilities
• Provides structured, type-safe outputs
• Includes comprehensive test coverage

Key Features

• Type-safe AI interactions using Pydantic models
• Structured dependency injection with dataclasses
• Async support for API operations
• OpenAI GPT-4 Vision integration
• Tool-augmented AI agent capabilities

Code Structure

The main components include:

• MultimodalLLMService: Service for interacting with OpenAI's vision models
• InvoiceProcessingDependencies: Dependency container for the AI agent
• InvoiceExtractionResult: Structured output model for extracted data
• Custom tools for processing invoice images

Prerequisites

Before running this project, ensure you have the following prerequisites:

• Python 3.8 or higher
• An OpenAI API key
• Required Python packages (listed in requirements.txt)

To install the required packages, run:

pip install -r requirements.txt

Running the Project

To run the project, use the following command:

python3 app.py

Testing

To run the tests, use the following command:

pytest