Getting Started

GitHub
Downloads

Frameworks

.NET

.NET MAUI

Operating Systems

Android

iOS

macOS

Shiny.DocumentIntelligence is two halves of one pipeline:

1. IDocumentScanner — launches the platform’s own document camera (edge detection, perspective correction, multi-page) and hands back clean page images.
2. IDocumentExtractor — turns those images into typed data: receipts, invoices, driver’s licenses, passports and payment cards.

Everything runs on-device. No cloud OCR service, no per-page cost, no document leaving the phone.

No MAUI dependency

This package gets its platform context natively — iOS walks connected scenes for the top view controller, Android tracks the current activity through a zero-config content provider. It works in a MAUI app, a native Android or iOS app, and a macOS AppKit app alike.

Install & register

dotnet add package Shiny.DocumentIntelligence

builder.Services.AddDocumentIntelligence();

That single call registers IDocumentScanner, ITextRecognizer, IBarcodeReader, IDataDetector and IDocumentExtractor — the native implementation for the current platform in each case.

Platform manifest entries

Platform	Required
iOS / Mac Catalyst	NSCameraUsageDescription in Info.plist
Android	Nothing — see below
macOS (AppKit)	Nothing — the user picks files through NSOpenPanel

Android needs no CAMERA permission

Do not add android.permission.CAMERA for this package. The ML Kit document scanner presents its own UI inside the Google Play services process, under that process’ camera permission — your app never touches the camera, and users choose which pages come back. Google describes it as “a complete solution eliminating the need for camera permission”. Requesting it anyway costs you a permission on your Play listing for nothing.

iOS is the opposite case: VisionKit’s document camera runs in your process, so the Info.plist entry is mandatory and the app is terminated without it.

Or generate the full set of files — packages, registration, manifests and permissions — with the builder below:

• NuGet Packages
• Project File
• MauiProgram.cs
• AndroidManifest.xml
• Info.plist
• PrivacyInfo.xcprivacy

Shiny.DocumentIntelligence

Scan and extract

public class ExpenseService(IDocumentScanner scanner, IDocumentExtractor extractor)
{
    public async Task<ReceiptData?> CaptureReceipt()
    {
        if (!scanner.IsSupported)
            return null;

        var scan = await scanner.ScanAsync(new DocumentScanRequest { PageLimit = 1 });
        if (scan.IsCancelled)
            return null;

        var doc = await extractor.ExtractAsync(scan, DocumentType.Receipt);
        return doc.Receipt;   // Merchant, Date, Subtotal, Tax, Total, Currency, Items
    }
}

Platform support

Platform	Scanner	OCR	Barcodes	Data detector
iOS / Mac Catalyst	VisionKit document camera	Apple Vision	Apple Vision	NSDataDetector
Android	ML Kit document scanner (can emit PDF)	ML Kit	ML Kit	—
macOS (AppKit)	File picker + Vision deskew	Apple Vision	Apple Vision	NSDataDetector
Windows / bare net10.0	throws	throws	throws	inert

Gate your UI on IsSupported rather than on a platform check — it also covers OS-version floors (the iOS document camera needs iOS 13+).

What can be extracted

DocumentType	Strategy	Result
Receipt	OCR + heuristics	ReceiptData — merchant, date, subtotal, tax, total, currency, line items
Invoice	OCR + heuristics	InvoiceData — vendor, number, dates, totals, line items
DriversLicense	AAMVA PDF417 barcode decode	LicenseData — name, DOB, licence number, address, plus the full AAMVA element map
Passport	ICAO 9303 MRZ parse	PassportData — names, passport number, nationality, dates, with check-digit validation
CreditCard	OCR + Luhn validation	CreditCardData — PAN, network, expiry, cardholder
Unknown	OCR only	RawText

See Extraction for the details, including the deliberate security choices in the payment-card path.

Where next

• Scanning — the modal scanner, per-platform behaviour, PDFs
• Extraction — typed results, OCR layout, the seams you can replace