Skip to content GitHub
Guide contents
English View source

OCR and Image Viewer

Visor & OCR Inspection

CyberPaste includes an adaptive media visor with local OCR text extraction, allowing you to inspect all clip types (images, text, code, rich text, HTML, URLs, and files) with full zoom, pan, and metadata tools without leaving the app.

Universal Media Visor

Opening the Visor

  • Right-click any clip → View (preview modal) or Open full viewer (full-screen window)
  • Double-click or press Enter on content cards where configured
  • Extract text (OCR) on image clips opens the viewer directly with the OCR drawer

Visor Features

Feature Description
Adaptive content inspection Universal viewer displaying high-resolution images, styled text, syntax-highlighted code, rendered HTML/RTF, link targets, and file icons
VISOR badge & metadata footer Header VISOR indicator and compact footer showing clip type, originating source app, and capture timestamps
Flash-free reveal Window opens hidden and displays only after theme styles load, eliminating white flashes
Zoom & Pan Mouse wheel or toolbar buttons (5% – 800%) for media, with drag-to-pan when enlarged
Fit to window Auto-fits content when opened; double-click toggles between fit and 1:1 scale
Copy & Export One-click copy or Save As export via native Windows file dialogs
OCR drawer Embedded panel with extracted text from images (copy / save / close)
Window memory Size, position, and maximized state are remembered across sessions with screen edge validation
Title-bar menu Full window control menu with theme awareness

The visor can save the displayed content directly. External image editing operations still use the configured external editor.

External Editor

Configure an external image editor:

  1. Go to Settings → General
  2. Set Image editor path to your preferred editor
  3. Right-click an image clip → Open external viewer

OCR Text Extraction

How OCR Works

  1. Right-click an image clip → Extract text (OCR) (or open the OCR drawer from the viewer)
  2. CyberPaste uses the Windows.Media.Ocr API; fully local, nothing leaves your machine
  3. The image is preprocessed and upscaled before recognition for better accuracy
  4. Text is reconstructed with layout awareness (line grouping and indentation estimation)
  5. Results appear in the viewer's OCR drawer and can be opened as an editable result dialog

Smart Post-Processing

  • Indent estimation — Median character width is used to convert pixel indents into spaces
  • Spanish diacritics correction — Common missing-accent patterns are fixed automatically

OCR Result Dialog

  • Extracted text in an editable text box
  • Copy button
  • Save As button (exports the OCR result as a text file)
  • Save to clip button (creates a new text clip)
  • Auto-copy toggle (persisted per installation; on by default)

ℹ️ The dialog does not show detected language or confidence scores.

Supported Languages

Windows Media OCR works with the languages installed on your system:

  • CyberPaste requests your chosen language and falls back to the user profile languages
  • Coverage depends on your Windows language packs

OCR Tips

  1. High resolution — Larger images give better results (upscaling is applied automatically)
  2. Good contrast — Dark text on light background works best
  3. Straight orientation — Avoid skewed or rotated text

Image Storage

Capture

Images are captured from the clipboard:

  • Full image bytes stored in the SQLite database
  • Thumbnails generated for the grid
  • Full files persisted under %APPDATA%\CyberPaste\images and cleaned up when clips are pruned
← All documentation