Modaic AutoAgent Quickstart

A simple quickstart project demonstrating how to use Modaic's AutoAgent for receipt classification and cost extraction from images.

Overview

This project shows how to:

• Use a precompiled agent from the Modaic Hub
• Extract cost information from receipt images
• Configure agent parameters like max tokens

Prerequisites

• Python 3.11 or higher
• Modaic account and API token
• OpenAI API key

Installation

1. Install dependencies:

   uv add modaic
   # or with pip:
   pip install modaic
   

2. Set up environment variables: Create a .env file in the project root with your API keys:

   MODAIC_TOKEN=your_modaic_token_here
   OPENAI_API_KEY=your_openai_api_key_here
   

Usage

Basic Example

from modaic import AutoAgent

# Load a precompiled receipt classifier from the Modaic Hub
receipt_classifier = AutoAgent.from_precompiled(
    "farouk1/receipt-classifier",
    config_options={"max_tokens": 500}
)

# Analyze a receipt image and extract the total cost
results = receipt_classifier(
    "https://ocr.space/Content/Images/receipt-ocr-original.webp"
)

print(results.total_cost)  # Output: 98.21

What This Does

1. Loads a precompiled agent - The farouk1/receipt-classifier agent is downloaded from the Modaic Hub
2. Configures the agent - Sets max_tokens to 500 for response length control
3. Processes an image - Analyzes the receipt image from the provided URL
4. Extracts cost data - Returns the total cost found on the receipt

Key Features

• Precompiled Agents: Use ready-made agents from the Modaic Hub
• Image Processing: Direct support for image URLs and analysis
• Configurable Parameters: Customize agent behavior with config options
• Simple API: Just a few lines of code to get started

Getting API Keys

Modaic Token

1. Sign up at Modaic
2. Navigate to your account settings
3. Generate an API token

OpenAI API Key

1. Sign up at OpenAI
2. Go to API keys section
3. Create a new API key

Project Structure

quickstart/
    .env                 # API keys (not committed to git)
    .gitignore          # Git ignore patterns
    main.py             # Agent implementation and hub upload
    pyproject.toml      # Project dependencies and metadata
    README.md           # This file
    uv.lock            # Dependency lock file

Running the Example

python main.py

This will run the quickstart example and display the extracted total cost from the receipt image.

Next Steps

• Explore more precompiled agents on the Modaic Hub
• Learn about creating your own agents with the Modaic SDK
• Check out the Modaic documentation for advanced features

Support

For questions or issues:

• Visit the Modaic documentation
• Check the GitHub repository
• Join the community discussions