🧾 Invoice Extraction Demo with Vision & Foundation Models

This is a lightweight iOS demo that shows how to extract structured data from an invoice image using the power of:

• 🔍 Apple's Vision framework for text recognition
• 🧠 Foundation Models for parsing structured data using on-device LLMs
• ✅ Safe and strongly typed output using @Generable and @Guide

📸 What It Does

This app demonstrates the end-to-end pipeline:

1. You provide or capture an image of an invoice
2. The app uses Vision to extract the printed text from the image
3. That raw text is sent into Apple’s on-device Foundation Model
4. The model returns structured data representing the invoice, using Swift types

📦 Output Model

The structured output is defined using the @Generable macro and @Guide descriptions to guide the LLM:

@Generable
struct InvoiceItem {
    var name: String
    var price: Decimal
    var quantity: Int
}

@Generable
struct MyInvoice {
    @Guide(description: "The name of the vendor")
    var vendor: String

    @Guide(description: "List of the invoice items")
    var items: [InvoiceItem]

    @Guide(description: "total invoice amount")
    var totalAmount: Decimal

    var toString: String {
        "Vendor: \(vendor)\n" +
        "Items:\n" +
        items.map(\.name).joined(separator: "\n") +
        "------\n" +
        "\nTotal: \(totalAmount)"
    }
}