Crate visionkit 

visionkit-rs

Safe Rust bindings for Apple’s VisionKit.framework on macOS.

Status: v0.2.0 covers ImageAnalyzer, ImageAnalysis, a headless-friendly LiveTextInteraction wrapper over macOS ImageAnalysisOverlayView, and ships explicit availability metadata for the iOS-only document-camera / Data Scanner / recognized-item areas.

Quick start

use visionkit::prelude::*;

fn main() -> Result<(), Box<dyn std::error::Error>> {
    if !ImageAnalyzer::is_supported() {
        println!("ImageAnalyzer is not supported on this Mac");
        return Ok(());
    }

    let analyzer = ImageAnalyzer::new()?;
    let configuration = ImageAnalyzerConfiguration::new(ImageAnalysisTypes::TEXT)
        .with_locales(["en-US"]);
    let analysis = analyzer.analyze_image_at_path(
        "examples/assets/live_text.png",
        ImageOrientation::Up,
        &configuration,
    )?;

    let interaction = LiveTextInteraction::new()?;
    interaction.track_image_at_path("examples/assets/live_text.png")?;
    interaction.set_analysis(&analysis)?;

    println!("transcript: {}", analysis.transcript()?);
    println!("live text overlay text: {}", interaction.text()?);
    Ok(())
}

Highlights

• ImageAnalyzer::is_supported and supported_text_recognition_languages
• ImageAnalyzerConfiguration, ImageAnalysisTypes, and ImageOrientation
• File-path based analysis through every public macOS analyzer overload: URL, NSImage, CGImage, CIImage, and CVPixelBuffer
• ImageAnalysis::transcript and has_results
• LiveTextInteraction control over macOS ImageAnalysisOverlayView state and inspection
• VNDocumentCameraViewController, DataScannerViewController, RecognizedText, Barcode, and RecognizedItem availability metadata on macOS

Availability

• ImageAnalyzer, ImageAnalysis, and LiveTextInteraction are available on macOS 13+.
• LiveTextInteraction::text and selected_text require macOS 14+ because Apple introduced those overlay accessors after macOS 13.
• VNDocumentCameraViewController, DataScannerViewController, RecognizedText, Barcode, and RecognizedItem are iOS-only Apple APIs. On macOS, this crate exposes them as structured availability metadata instead of pretending they exist.

Examples

cargo run --example 01_vn_document_camera_view_controller
cargo run --example 02_framework_smoke
cargo run --example 03_data_scanner_view_controller
cargo run --example 04_image_analyzer
cargo run --example 05_live_text_interaction
cargo run --example 06_image_analysis
cargo run --example 07_recognized_text
cargo run --example 08_barcode
cargo run --example 09_recognized_item

See COVERAGE.md for the audited VisionKit API matrix.

License

Licensed under either of Apache-2.0 or MIT at your option.

Re-exports

pub use barcode::Barcode;

pub use data_scanner_view_controller::DataScannerViewController;

pub use error::VisionKitError;

pub use image_analysis::ImageAnalysis;

pub use image_analyzer::ImageAnalysisTypes;

pub use image_analyzer::ImageAnalyzer;

pub use image_analyzer::ImageAnalyzerConfiguration;

pub use image_analyzer::ImageOrientation;

pub use live_text_interaction::EdgeInsets;

pub use live_text_interaction::LiveTextInteraction;

pub use live_text_interaction::LiveTextInteractionTypes;

pub use live_text_interaction::Rect;

pub use recognized_item::RecognizedItem;

pub use recognized_text::RecognizedText;

pub use support::AreaSupportInfo;

pub use vn_document_camera_view_controller::VNDocumentCameraViewController;

Modules

analyzer

barcode

data_scanner_view_controller

error

ffi

image_analysis

image_analyzer

live_text_interaction

prelude

recognized_item

recognized_text

support

vn_document_camera_view_controller