vision-framework

Extract text, detect faces and barcodes from iOS photos and video frames.

Updated Apr 19, 2026

Search Tags:#ocr #swift #vision #face-detection #barcode-detection #document-scanning

Ready to run

npx skills add https://github.com/femitz/flyby --skill vision-framework-femitz

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: vision-framework
Source: https://github.com/femitz/flyby/tree/main/.agents/skills/vision-framework
Command: npx skills add https://github.com/femitz/flyby --skill vision-framework-femitz

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Implements computer vision capabilities in iOS apps using Vision and VisionKit to perform OCR, face and barcode detection, document scanning, and segmentation, all on-device.

Core Features & Use Cases

On-device text recognition (OCR) with multilingual support and language hints.
Real-time face detection with landmarks, body pose, and quality metrics.
Barcode detection across multiple symbologies with payload extraction.
Document scanning and structured data extraction for downstream OCR pipelines.
VisionKit integration for live scanning and SwiftUI overlays.

Quick Start

Provide an image and ask the system to perform OCR to extract all readable text.

Frequently Asked Questions about vision-framework

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I perform on-device OCR text recognition in iOS using Swift?▼

On-device OCR text recognition in iOS uses the Vision framework to extract readable text from images or video frames. It supports multilingual content with language hints and operates asynchronously to avoid blocking the main thread.

Can I detect faces and barcodes simultaneously in real-time video frames?▼

Face detection and barcode detection can process real-time video frames using Vision APIs. Face detection includes landmarks, body pose, and quality metrics, while barcode detection extracts payloads across multiple symbologies.

Does the iOS Vision framework support legacy platforms older than iOS 18?▼

Modern iOS 18+ Vision APIs are supported with fallbacks for legacy platforms. This ensures robust operation and non-blocking threading across different device versions using asynchronous perform(on:) workflows.

What's the best way to scan documents and extract structured data on iOS?▼

Document scanning and structured data extraction use VisionKit integration for live scanning. This provides a pipeline that feeds scanned document images directly into OCR for downstream text extraction.

How do I convert Vision face detection coordinates for SwiftUI overlays?▼

Coordinate conversions for UI overlays translate Vision framework detection results into SwiftUI coordinate spaces. This allows accurate mapping of detected face landmarks and bounding boxes onto visual interface layers.

Why does my Vision OCR request fail to recognize non-English text?▼

OCR requests failing on non-English text often lack proper language hints. Specifying language hints in the Vision request configuration enables multilingual text recognition for accurate extraction.