Text Lens provides universal text extraction from any screen content using on-device OCR processing. The app converts non-selectable text from images, videos, PDFs, and presentations into editable text through a simple drag-to-select interface.
The workflow uses a customizable global shortcut to activate capture mode. Users drag to select any screen region, and Text Lens immediately recognizes and extracts all text within that area. Recognition accuracy handles blurry screenshots, technical documentation, and stylized fonts across 18+ languages.
Key features include:
- Screen region capture with instant OCR text extraction
- Clipboard image text extraction
- Direct file scanning for images and documents
- Camera scanning for physical documents
- QR code and barcode recognition
- Support for 18+ languages including stylized fonts
Processing occurs entirely on-device using Apple’s Vision framework. No cloud services, data collection, or account creation is required. This local-only approach makes the app suitable for sensitive documents and regulated environments.
Performance remains responsive with sub-second recognition for most content. Memory usage stays around 50MB when active and minimal when idle. The menu bar integration provides clean access to capture mode.
System requirements include macOS 14 or newer. Text Lens costs $6 on the Mac App Store. The app is developed by Sindre Sorhus.
Limitations include reduced accuracy with very stylized or artistic fonts, a constraint affecting all OCR tools. Standard text on screens and documents achieves high extraction accuracy.
Alternatives include macOS built-in Live Text for basic OCR within supported apps, CleanShot X for screenshot tools with OCR capabilities, or TextSniper for similar screen OCR functionality.
Suitable for users who regularly encounter non-selectable text, including students, researchers, developers, and professionals who need to extract text from screenshots, video tutorials, or scanned documents while maintaining privacy through local processing.
