Repository navigation
Commander v1.1.0 - Screen Interaction & New Models
Major update with AI-powered screen interaction and latest models from OpenAI, Google, and Anthropic.
What's New
Screen Interaction Tools
- type_text: AI can type into any focused input field (Gmail, Notes, code editors, etc.)
- click_element: AI can click buttons, links, and UI elements
- find_ui_elements: AI discovers interactive elements using accessibility APIs
- Smart Confirmation: Destructive actions (delete, send, pay, logout) require approval
Vision Improvements
- Better screenshot capture with improved frame validation
- Graceful fallback to full display capture
- Fixed "Invalid frame dimension" errors
- Improved window filtering (aspect ratio, minimum dimensions)
New AI Models
OpenAI:
- GPT-5.2, GPT-5.2-pro (latest flagship)
- GPT-5-mini, GPT-5-nano
- o3-pro (enhanced reasoning)
Google:
- Gemini 3 Flash Preview
- Gemini 3 Pro Preview
AI Identity
- Meet Jarvis, your personal AI assistant
Bug Fixes
- Fixed tool execution loop (prevented infinite tool calls)
- Improved screen capture reliability
- Better error handling throughout
Installation
- Download
Commander-v1.0.dmg - Open the DMG and drag Commander to Applications
- Launch Commander and grant required permissions:
- Accessibility (for hotkeys & screen interaction)
- Microphone (for voice input)
- Screen Recording (for vision features)
Requirements
- macOS 13.0 (Ventura) or later
- Apple Silicon or Intel Mac
🤖 Generated with Claude Code