Act Desktop

Act Desktop is a desktop automation tool that brings agentic AI to non-technical users, enabling them to automate complex workflows across local files and web browsers.

Features

• Natural Language Tasks - Describe what you want to accomplish in plain English
• File Operations - Organize, rename, and transform files across your local folders
• Browser Automation - Extract data from websites and fill forms (via Nova Act)
• Human-in-the-Loop - Approve sensitive actions like file deletions and form submissions
• Privacy-First - All data stays local, you control your own API keys

Quick Start

1. Download the latest release from the Releases page
2. Install the application:
   • macOS: Open the .dmg and drag Act Desktop to Applications
   • Windows: Run the .exe installer
   • Linux: Use the .AppImage or .deb package
3. Set up API keys in the onboarding flow:
   • Get a Claude API key (required)
   • Get a Nova Act API key (optional, for browser automation)
4. Add folders you want the agent to access
5. Start automating!

Example Tasks

"Organize my Downloads folder by file type"

"Create a summary of all PDF files in Documents"

"Find all images larger than 5MB and list them"

"Extract data from a webpage and save to CSV"

Development

Prerequisites

• Node.js 18+
• npm 9+
• macOS, Windows, or Linux

Setup

# Install dependencies
npm install

# Start development server
npm run dev

# Run tests
npm test

# Build for production
npm run build

# Package for distribution
npm run build:mac   # macOS
npm run build:win   # Windows
npm run build:linux # Linux

Project Structure

act-desktop/
├── src/
│   ├── main/           # Electron main process
│   │   ├── agents/     # AI agent definitions
│   │   ├── tools/      # Agent tools (file ops, browser, etc.)
│   │   ├── services/   # Core services (credentials, sessions)
│   │   └── utils/      # Utilities (security, logging, errors)
│   ├── preload/        # Secure IPC bridge
│   └── renderer/       # React UI
│       └── components/ # UI components
├── tests/              # Test suites
└── resources/          # App icons and assets

Architecture

Act Desktop uses:

• Electron - Cross-platform desktop framework
• React + TypeScript - Modern UI with type safety
• Strands Agents SDK - AI agent orchestration
• Claude API - Anthropic's Claude for intelligence
• Nova Act SDK - Browser automation (optional)

┌─────────────────────────────────────────────┐
│           Electron Application              │
├─────────────────────────────────────────────┤
│  Main Process                               │
│  ├─ Orchestrator Agent (Claude)            │
│  ├─ File System Tools                      │
│  └─ Browser Tools (Nova Act)               │
├─────────────────────────────────────────────┤
│  Renderer Process (React UI)               │
│  ├─ Chat Interface                         │
│  ├─ Task Progress                          │
│  └─ Approval Dialogs                       │
└─────────────────────────────────────────────┘

Security

• Context Isolation - Renderer has no direct Node.js access
• Path Validation - File operations restricted to approved folders
• Secure Credentials - API keys stored in OS keychain
• Human Approval - Required for file deletions, form submissions, payments

Configuration

API keys are stored securely in your system keychain. Approved folders and preferences are stored locally.

Environment variables for development:

# .env.local
ANTHROPIC_API_KEY=your-api-key
NOVA_ACT_API_KEY=your-nova-act-key  # Optional

Contributing

Contributions are welcome! Please read our contributing guidelines before submitting PRs.

License

MIT License - see LICENSE for details.

Acknowledgments

• Anthropic for Claude
• Electron team
• shadcn/ui for UI components