Intelligent Windows automation agent powered by Claude AI with beautiful Flutter GUI
Transform natural language commands into Windows actions: install apps, automate tasks, browse the web, download images, and control your desktop—all through an elegant chat interface.
- Install applications from downloads folder with intelligent automation
- Check installed software and versions
- Uninstall applications by name
- Launch and control apps with UI automation
- Browser control with native Chrome profile support
- Google Images search & download - bulk image downloads
- Form filling and web navigation
- Download monitoring with OneDrive compatibility
- File operations - create, move, copy, search files
- Process management - kill, list, monitor processes
- System info - hardware details, disk space, network
- Desktop automation - keyboard/mouse control, screenshots
- OCR text extraction with Tesseract
- Installer automation - Python, Chrome, etc with V2 robust detection
- Cross-PC compatibility - OneDrive path handling
- 77 automation tools available through Claude API
- 🎨 Frameless modern design with purple/cyan gradient theme
- 💬 Chat interface with message history
- 🔧 Tool execution tracking with success/error badges
- 🎛️ Model selection - Switch between Haiku, Sonnet, Opus
- 🟢 Live connection status indicator
# Check Python version
python --version # Should be 3.12 or higher
# If not installed, download from:
# https://www.python.org/downloads/# Check Flutter version
flutter --version # Should be 3.35.3 or higher
# If not installed:
# 1. Download Flutter SDK: https://docs.flutter.dev/get-started/install/windows
# 2. Extract to C:\flutter
# 3. Add to PATH: C:\flutter\bin
# 4. Run: flutter doctorGet your API key from Anthropic Console
# Clone the repository
git clone https://github.com/bxf1001g/axonyx-revolt.git
cd axonyx-revolt
# Run automated setup
.\start.ps1This automatically:
- ✅ Creates Python virtual environment
- ✅ Installs Python dependencies
- ✅ Creates
.envfile from template - ✅ Launches the CLI agent
# 1. Clone repository
git clone https://github.com/bxf1001g/axonyx-revolt.git
cd axonyx-revolt
# 2. Create virtual environment
python -m venv .venv
.\.venv\Scripts\Activate.ps1
# 3. Install Python dependencies
pip install -r requirements.txt
# 4. Setup environment
Copy-Item .env.example .env
# Edit .env and add your ANTHROPIC_API_KEY
# 5. Run CLI agent
python src/main.py# Run automated GUI setup
.\setup_gui.ps1This will:
- ✅ Verify Flutter installation
- ✅ Check Python virtual environment
- ✅ Install FastAPI and Uvicorn
- ✅ Download Flutter dependencies
- ✅ Enable Windows desktop support
- ✅ Create Windows runner files
# Option 1: Launch both (Recommended)
.\launch_gui.ps1 -Both
# Option 2: Launch backend only
.\launch_gui.ps1 -Backend
# Option 3: Launch GUI only (backend must be running)
.\launch_gui.ps1 -GUIThe GUI will open at 1200x800 frameless window with:
- Dark theme with purple/cyan gradients
- Chat interface with message bubbles
- Model selector dropdown
- Tool execution badges
- Live backend connection indicator
# Required
ANTHROPIC_API_KEY=sk-ant-api03-your-key-here
# Model Selection (choose one)
CLAUDE_MODEL=claude-3-5-haiku-20241022 # Fast & Cheap (Recommended)
# CLAUDE_MODEL=claude-3-7-sonnet-20250219 # Balanced
# CLAUDE_MODEL=claude-opus-4-20250514 # Most Powerful
# Optional Settings
REQUIRE_CONFIRMATION=false # GUI mode doesn't need confirmation
MAX_ITERATIONS=10
DRY_RUN_MODE=false| Model | Cost (Input/Output) | Speed | Best For |
|---|---|---|---|
| Haiku 3.5 | $0.80 / $4 per MTok | ⚡⚡⚡ Fastest | Automation, quick tasks |
| Sonnet 3.7 | $3 / $15 per MTok | ⚡⚡ Fast | Complex reasoning |
| Opus 4 | $15 / $75 per MTok | ⚡ Moderate | Most difficult tasks |
💡 Recommendation: Start with Haiku 3.5 for automation tasks.
axonyx-revolt/
├── src/
│ ├── agent.py # Core agent with 77 tools
│ ├── api_server.py # FastAPI backend for GUI
│ ├── main.py # CLI interface
│ └── tools/ # Automation tool modules
│ ├── file_ops.py # File operations (8 tools)
│ ├── process_ops.py # Process management (6 tools)
│ ├── ui_automation.py # Desktop automation (10 tools)
│ ├── browser_automation.py # Web control (12 tools)
│ ├── chrome_launcher.py # Native Chrome (2 tools)
│ ├── image_downloader.py # Image downloads (3 tools)
│ ├── screen_reader.py # OCR tools (6 tools)
│ ├── installer_automation.py # Basic installers (5 tools)
│ ├── installer_automation_v2.py # Robust installers (5 tools)
│ ├── download_manager.py # Download tracking (4 tools)
│ ├── app_installation.py # App management (5 tools)
│ ├── app_control.py # App control (4 tools)
│ ├── system_info.py # System info (4 tools)
│ └── system_settings.py # Settings control (3 tools)
├── axonyx_gui/ # Flutter Windows GUI
│ ├── lib/
│ │ ├── main.dart # App entry point
│ │ ├── theme/ # Dark theme with gradients
│ │ ├── models/ # Data models
│ │ ├── providers/ # State management
│ │ ├── screens/ # Chat screen
│ │ └── widgets/ # Reusable UI components
│ └── windows/ # Windows desktop runner
├── docs/ # Comprehensive documentation
├── examples/ # Usage examples
├── requirements.txt # Python dependencies
├── .env.example # Environment template
├── setup_gui.ps1 # GUI setup script
├── launch_gui.ps1 # GUI launcher
└── start.ps1 # CLI launcher
**77 Tools Total** across 13 modules
python src/main.pyExample commands:
> Install Python from my Downloads folder
> Take a screenshot of my desktop
> Open Chrome and search for "AI automation"
> List all running processes using more than 100MB RAM
> Download 10 images of sports cars
> Create a new folder called "Projects" on my Desktop
- Launch with
.\launch_gui.ps1 -Both - Select model from dropdown (Haiku 3.5 recommended)
- Type natural language commands in the input field
- Watch tool execution with badges:
- ✅ Success
- ❌ Error
- 🔄 In progress
Command: "Download 10 images of cars"
Process:
1. Opens Chrome to Google Images
2. Searches for query
3. Clicks thumbnails systematically
4. Downloads full-size images
5. Saves to: Downloads/google_images/cars/
Result: 10 JPG files downloaded
Robust multi-strategy installer automation:
- Exact window title matching
- Fallback strategies (4 levels)
- Checkbox detection with multiple methods
- Detailed step-by-step results
- Works with: Python, Chrome, Node.js, Git, etc.
Automatically handles OneDrive-redirected paths:
Desktop → C:\Users\USERNAME\OneDrive\Desktop
- GETTING_STARTED.md - First-time setup
- SETUP_GUIDE.md - Detailed installation
- docs/FLUTTER_GUI_SETUP.md - GUI setup guide
- MODEL_SELECTION_GUIDE.md - Choose the right model
- CONFIRMATION_GUIDE.md - Confirmation mode
- CROSS_PC_COMPATIBILITY.md - Multi-PC setup
- TESSERACT_SETUP.md - OCR setup
- HYBRID_INSTALLER_AUTOMATION.md - Installer automation
- docs/IMAGE_DOWNLOAD_FEATURE.md - Image downloads
- docs/INSTALLER_AUTOMATION_V2.md - V2 installers
- examples/ - Code examples
- QUICK_REFERENCE.md - One-page cheat sheet
- PROJECT_OVERVIEW.md - Architecture overview
- docs/CORRECT_MODEL_IDS.md - Model IDs
- ✅ API keys stored locally in
.envfile (not committed to git) - ✅ No data sent to third parties except Anthropic API
- ✅ Confirmation mode available for safety
- ✅ OneDrive-aware path handling
⚠️ Use with caution - agent has system-level access
Contributions are welcome! Please:
- Fork the repository
- Create a feature branch (
git checkout -b feature/amazing-feature) - Commit your changes (
git commit -m 'Add amazing feature') - Push to the branch (
git push origin feature/amazing-feature) - Open a Pull Request
This project is licensed under the MIT License - see the LICENSE file for details.
# Ensure virtual environment is activated
.\.venv\Scripts\Activate.ps1
# Reinstall dependencies
pip install -r requirements.txt# Add Flutter to PATH
$env:PATH += ";C:\flutter\bin"
# Verify
flutter doctor# Kill existing process
netstat -ano | findstr :8000
Stop-Process -Id <PID> -ForceCheck that you're using correct model IDs:
- ✅
claude-3-5-haiku-20241022 - ✅
claude-3-7-sonnet-20250219 - ✅
claude-opus-4-20250514
See docs/CORRECT_MODEL_IDS.md for details.
- Anthropic - Claude AI API
- Flutter - Beautiful GUI framework
- Python - Automation backbone
- pywinauto - Windows UI automation
- Selenium - Browser automation
- Issues: GitHub Issues
- Discussions: GitHub Discussions
Made with ❤️ for Windows automation enthusiasts
⭐ Star this repo if you find it useful! ⭐