Created by Arian Rudd

Turn ideas into production-ready apps and solutions. Zero code required. CODAI enables anyone to build fully functional applications using natural language.

🌐 Website: codai.ai
🐦 Creator: @AriRudd

CODAI/
├── server/              # FastAPI backend
│   ├── routes/         # API endpoints
│   ├── middleware/     # Request middleware
│   └── app.py         # Main application
├── frontend/           # Next.js frontend
│   ├── src/           # Source code
│   │   ├── components/# React components
│   │   ├── hooks/     # Custom hooks
│   │   └── lib/       # Utilities
│   └── package.json   # Frontend dependencies
├── core/              # Core backend functionality
│   ├── api/          # Claude API integration
│   ├── logging/      # Logging system
│   └── schemas/      # Data schemas
├── tools/            # Tool implementations
│   ├── computer/     # Computer use tools
│   ├── web/          # Web interaction
│   └── pdf/          # PDF processing
├── config/           # YAML configuration
└── Docs/             # Documentation

• Python 3.9+
• Node.js 18+
• Git
• Anthropic API key

The open-source web version allows you to run CODAI in your browser with full functionality. Perfect for:

• Immediate access without installation
• Cross-platform compatibility (Windows, Mac, Linux)
• Team collaboration
• Cloud deployment

Native Electron desktop application with enhanced capabilities:

• Self-contained application (no separate install required)
• Local file system access with native dialogs
• Bundled Python backend (no Python install needed)
• Works offline once installed
• Available for Windows, macOS, and Linux

1. Clone the repository:

git clone https://github.com/thecodai/Codai
cd Codai

2. Choose your platform:

# Run the Electron setup wizard
.\setup-electron.ps1
# or
setup-electron.bat

# Start the desktop app
npm run dev

# Run the Electron setup
./setup-electron.sh

# Start the desktop app
npm run dev

.\setup.ps1

This provides a user-friendly wizard interface with colored output and system checks.

setup.bat

python -m setup

./setup.sh

This provides a user-friendly wizard interface with colored output and system checks.

python -m setup

This will:

• Create virtual environment
• Install Python dependencies
• Install frontend dependencies
• Set up configuration files
• Initialize the database

4. Set your API key in the UI:

• Open http://localhost:8001
• Click Settings → API Keys
• Add your Anthropic API key

python run_server.py
# Or with options:
python run_server.py --env development --log-level debug

cd frontend
npm run dev

# Backend tests
pytest

# Frontend tests
cd frontend
npm test
npm run lint

• No Coding Required: If you can describe it, CODAI can build it
• As Simple as Google: Just type what you want in plain English
• Complete Applications: Not just code snippets - full, production-ready apps
• Open Source: Web-based version available now, desktop version coming soon

• Claude 4 Integration: Latest Anthropic models with extended thinking
• Intelligent Context Management: Automatic summarization and token optimization
• Multi-Model Support: Claude 4 Opus, Sonnet, and Haiku models
• Computer Use: Automated screen capture and interaction capabilities

• Secure API Key Management: Military-grade encryption (AES-256-GCM)
• Session Management: Persistent conversations with import/export
• Comprehensive Monitoring: Real-time performance and usage analytics
• Error Recovery: Robust error handling with automatic recovery

# Encryption secret (auto-generated on first run if not set)
ENCRYPTION_SECRET=<auto-generated-or-your-custom-secret>

# Optional - users can add via UI instead
ANTHROPIC_API_KEY=sk-ant-api-key

# Environment
APP_ENV=development

Note: The encryption secret is now automatically generated on first server start if not provided.

• config/base.yaml - Base configuration
• config/environments/*.yaml - Environment overrides
• config/local.yaml - Local overrides (git ignored)

Users can add their Anthropic API keys through the UI:

1. Open Settings (gear icon)
2. Navigate to API Keys
3. Enter your key (must start with sk-ant-)
4. Save and test the connection
5. Click "Continue to Chat" when validated

Keys are:

• Encrypted with AES-256-GCM
• Stored securely in the backend database
• Never exposed in logs or responses
• Tied to user sessions

Configure complex reasoning in config/base.yaml:

ai:
  extended_thinking:
    enabled: true
    budget_tokens: 5000  # 1024-16000
    show_thinking: true
    auto_detect: true

• API Keys: Encrypted storage with session isolation
• CORS: Configured for development/production
• Rate Limiting: 60 requests/minute per IP
• Input Validation: Comprehensive request validation
• Audit Logging: All API key operations logged

See /Docs/New/ for detailed documentation:

• Architecture overview
• Feature implementations
• API documentation
• Configuration guide

This project is licensed under the GNU General Public License v3.0 - see the LICENSE file for details.

Copyright (c) 2025 Arian Rudd

The author kindly requests that any derivative works or implementations:

1. Keep the CODAI branding or mention "Based on CODAI" in your derivative works
2. Include attribution: "Originally created by Arian Rudd"
3. Link back to this repository when practical

The CODAI logo and visual identity are proprietary. Please see BRANDING.md for usage guidelines.

Under GPL-3.0, you must:

• Provide source code when distributing the software
• License any derivative works under GPL-3.0
• Include copyright and license notices
• Document any changes made to the original code

This program is distributed in the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for more details.

Created and maintained by Arian Rudd

CODAI represents a significant investment of time and effort. If you find it useful, please consider:

• ⭐ Starring this repository
• 🔗 Linking back to this project
• 📣 Sharing your experience with CODAI
• 🌐 Visiting codai.ai for more information

• Website: codai.ai
• X (Twitter): @AriRudd
• GitHub: thecodai/Codai

For questions about commercial use or custom implementations, please visit codai.ai or reach out on X.