- Migrate Swift code to Swift 6 compatibility
- Rename swift-cli folder to peekaboo-cli for better naming consistency
- Update all references throughout the codebase (build scripts, imports, tests)
- Fix .gitignore to properly track Swift source files
- Update environment variable names to use PEEKABOO_ prefix
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Configure CI to run on macOS-latest
- Test with Node.js 20.x and 22.x
- Run npm build and tests on push/PR
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add humorous dialogue showing the pain of explaining visual bugs to blind AI
- List practical benefits: bug hunting, design reviews, UI testing, etc.
- Position Peekaboo as "supernatural contact lenses" for AI assistants
- Keep the spooky theme while explaining real developer value
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Add ghost emojis and supernatural terminology throughout
- Rename sections with mystical themes (Installation → Summoning, etc.)
- Transform technical terms into magical concepts
- Update all LOG_LEVEL references to lowercase format
- Maintain technical accuracy while adding Halloween personality
👻 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Convert to scoped npm package for easier installation
- Simplify installation to use npx (no global install needed)
- Update README to prioritize npm over AppleScript
- Add banner image with absolute GitHub URL
- Configure package for public npm publishing
Added automated tests to verify window capture functionality:
- Window bounds capture verification (checks -R flag usage)
- Window vs fullscreen size comparison
- Multi-window bounds capture testing
- Fallback message format verification
- Window capture with AI analysis
Also fixed fallback message to use Peekaboo prefix when falling back
to fullscreen capture due to inability to get window bounds.
Test additions:
- New "window" test mode for focused testing
- Added to quick and all test suites
- Comprehensive coverage of window capture edge cases
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Previously, window capture was falling back to fullscreen because it couldn't
get window IDs. Now using window bounds (position + size) with screencapture's
-R flag to properly capture only the app window content.
Changes:
- Use System Events to get window position and size
- Build screencapture -R command with x,y,width,height
- Added captureWindowByIndex for multi-window scenarios
- Proper error handling with fallback to fullscreen
This ensures AI analysis and screenshots only show the targeted app,
not the entire desktop.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Removing version references from README makes it more maintainable.
The version is tracked in the script itself and in git tags.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Since Peekaboo hasn't shipped yet, these aren't "new" features - they're
just the features of v1.0. Updated README to reflect this accurately.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Major improvements:
• 🪟 Smart Multi-Window AI: Automatically analyzes ALL windows when app has multiple
• ⏰ Timeout Protection: 90-second timeout prevents hanging on slow AI models
• 🎯 Clean CLI Design: Consistent flags (-o, -f, -w, -m, -a, -q, -v) with logical defaults
• 📝 All Messages Prefixed: "Peekaboo 👀:" for clear AI tool integration
• 🔄 Auto-Detection: Enables multi-window mode automatically for AI analysis
• 🎛️ Override Control: Use -w flag to force single window analysis
Technical changes:
- Changed version to 1.0.0 (initial release)
- Added formatMultiWindowAnalysis for consolidated AI results
- Improved error messages with proper context
- Enhanced test suite with multi-window AI tests
- Updated README with comprehensive documentation
This release provides a production-ready screenshot tool with intelligent
multi-window handling and robust AI integration.
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
🚀 Major documentation enhancement for AI tool integration:
📖 New Cloud AI Integration Section:
• Complete setup guide for Claude Code, Windsurf, ChatGPT, and other AI tools
• Simple 3-step process: copy file, tell AI about it, let AI auto-discover features
• Self-documenting design means AI tools instantly understand all capabilities
💡 Practical Integration Examples:
• Example prompts for different AI assistants
• Python integration code for custom AI scripts
• Real-world workflow examples combining cloud + local AI
🎯 Key Benefits Highlighted:
• Zero configuration required - just copy and use
• Smart filenames perfect for AI workflows
• Comprehensive feature auto-discovery by AI tools
• Reliable unattended operation ideal for AI automation
🔗 Enhanced Feature Coverage:
• Added Cloud AI Integration to main feature list
• Updated features table with new capabilities
• Emphasized self-documenting nature for AI consumption
This makes Peekaboo incredibly easy to integrate with any AI development workflow\!
🚀 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
✨ Major AI functionality improvements and documentation enhancements:
📏 Smart Image Compression:
• Auto-detects large images (>5MB) and compresses to 2048px for optimal AI processing
• Uses macOS native sips tool for quality resizing while preserving aspect ratio
• Preserves original screenshots untouched while creating temporary compressed versions
• Significantly improves AI analysis speed and reliability
🛠️ Enhanced Ollama Integration:
• Fixed AppleScript file writing to handle large JSON payloads (replaces shell limitations)
• Improved base64 encoding with newline removal for clean API calls
• Added robust JSON response parsing with proper escape handling
• Enhanced error handling for model availability and API connectivity
🎯 Argument Parsing Improvements:
• Fixed flag parsing to properly handle --verbose --ask combinations
• Added fullscreen mode support for --ask without app targeting
• Improved app identifier resolution with fullscreen handling
• Enhanced validation and error reporting
📖 Comprehensive README Enhancements:
• Highlighted smart image resizing feature with technical details
• Added complete step-by-step Ollama setup guide with multiple installation options
• Enhanced model comparison table with sizes, strengths, and use cases
• Added pro usage examples showcasing automatic compression benefits
• Improved formatting with clear sections and visual highlights
🧪 AI Testing Validation:
• All AI tests now passing with reliable Ollama API integration
• Verified one-step and two-step AI analysis workflows
• Confirmed smart compression works seamlessly with various image sizes
🚀 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Enhanced testing section to showcase the robust AI test suite:
New Testing Documentation:
• Highlighted 8 comprehensive AI vision analysis tests
• Added specific test mode examples (ai, advanced, basic, quick)
• Detailed AI test coverage breakdown
• Clear command examples for different testing scenarios
AI Test Coverage Highlights:
• One-step: Screenshot + AI analysis workflow
• Two-step: Analyze existing images separately
• Custom model specification and auto-detection
• Error handling: Invalid models, missing files, malformed commands
• Graceful degradation when Ollama/models unavailable
Testing Commands:
./test_peekaboo.sh ai # AI tests only
./test_peekaboo.sh advanced # Multi-window, discovery, AI
./test_peekaboo.sh all # Comprehensive suite
Perfect for developers wanting to validate AI functionality\!
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Enhanced test suite with full AI testing capabilities:
Test Coverage:
• One-step workflow: Screenshot + AI analysis in single command
• Two-step workflow: Analyze existing images separately
• Custom model testing with available vision models
• Error handling: Invalid models, missing files, malformed commands
• Smart model auto-detection validation
• Complex questions with special characters
Features:
• Auto-detects available Ollama installation and vision models
• Gracefully skips AI tests if Ollama/models not available
• Provides setup instructions when missing dependencies
• New dedicated 'ai' test mode for focused AI testing
• Integrated into advanced and comprehensive test suites
Examples:
./test_peekaboo.sh ai # AI tests only
./test_peekaboo.sh advanced # Includes AI tests
./test_peekaboo.sh all # Full suite with AI
Perfect for validating AI vision analysis functionality\!
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Major new feature: Local AI vision analysis with Ollama integration
Features:
• One-step: Screenshot + AI analysis in single command
• Two-step: Analyze existing images separately
• Smart model auto-detection with priority ranking
• Simplified ollama run commands (no complex API calls)
• Comprehensive error handling and setup instructions
Priority models: qwen2.5vl:7b > llava:7b > llava-phi3:3.8b > minicpm-v:8b
Examples:
osascript peekaboo.scpt "Safari" --ask "What's on this page?"
osascript peekaboo.scpt analyze "/tmp/shot.png" "Any errors?"
Perfect for automated testing, QA, and visual verification\!
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
Enhanced documentation to highlight new features:
• Added dedicated Smart Filenames section with examples
• Updated feature list with model-friendly naming & speed
• Enhanced examples throughout showing actual output filenames
• Updated "Why Peekaboo Rocks" with performance improvements
• Added specific timing details (0.3s, 0.2s, 0.1s delays)
• Emphasized automation and AI model compatibility
Now clearly communicates the value of recent optimizations\!
🤖 Generated with [Claude Code](https://claude.ai/code)
Co-Authored-By: Claude <noreply@anthropic.com>
