AIEXEcode is a CLI tool where you simply describe what you want in natural language, and AI handles the coding for you.

Just say things like "fix the lint errors in this file" or "add login functionality", and the AI will analyze the code, create a plan, and execute it. No complex commands or configurations needed - just speak naturally.

• 🐛 Bug Fixes: "Fix the error that occurs during login"
• ✨ Add Features: "Create a profile editing feature"
• 🔄 Refactoring: "Clean up auth.js to make it more readable"
• 🧪 Write Tests: "Write test code for all APIs"
• 📝 Documentation: "Add description comments to main functions"

Supported Operating Systems:

• macOS
• Linux

⚠️ Windows is not currently supported.

macOS:

brew install ripgrep

Ubuntu/Debian:

sudo apt-get install ripgrep

Fedora/CentOS:

sudo dnf install ripgrep

Arch Linux:

sudo pacman -S ripgrep

npm install -g aiexecode

Done! The aiexecode command is now available everywhere.

aiexecode

On first run, the setup wizard appears:

1. Enter API key
2. Select model

Where to get API keys:

• OpenAI: https://platform.openai.com/account/api-keys
• Anthropic: https://console.anthropic.com/settings/keys

Interactive Mode (Recommended):

aiexecode

A prompt appears, and you can enter your desired task:

> Refactor the user authentication module
> Fix all lint errors
> /exit

Quick Execution (Optional):

aiexecode "simple task"

That's all there is to it!

You can pause work and continue later:

# First start
aiexecode
> Create large file processing feature
# Output: New session ID: abc1234567890def

# Continue later
aiexecode -c abc1234567890def
> Now add error handling too

If you're curious what the AI did, you can view it in a web browser:

aiexecode --viewer

Open http://localhost:3300 in your browser to see:

• Commands executed by AI
• File modification history
• AI's thinking process

Convenient commands you can use during conversation:

/help          # Help
/exit          # Exit
/clear         # Clear screen
/apikey        # Change API key
/model         # Change model

AIEXEcode supports OpenAI GPT-5 series models:

• gpt-5
• gpt-5-mini (default)
• gpt-5-nano
• gpt-5-codex

Use /model list to see all available models or /model <model-name> to switch models.

If you want different AI behavior for each project:

aiexecode --init

A .aiexe/prompts/ folder is created where you can customize AI behavior.

If you need more powerful features, you can connect MCP servers:

# Connect GitHub
aiexecode mcp add --transport stdio github -- npx -y @modelcontextprotocol/server-github

# View connected servers
aiexecode mcp list

Required:

• macOS or Linux (Windows not supported)
• Node.js 14 or higher
• ripgrep (code search tool)
• OpenAI API key

Optional:

• Python 3 (for Python code execution support)

• Website: https://aiexecode.com
• Report bugs: https://github.com/kstost/aiexecode/issues

THIS SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED.

AIEXEcode is an experimental project. By using this software, you acknowledge and agree that:

• This tool is provided for experimental and educational purposes
• The developers and contributors are not liable for any damages, data loss, or issues arising from the use of this software
• Users assume all risks associated with using an AI-powered code modification tool
• Always review AI-generated code changes before committing them to production
• Maintain proper backups of your codebase before using this tool
• The software may produce unexpected results or behaviors

USE AT YOUR OWN RISK. The authors and contributors shall not be held responsible for any direct, indirect, incidental, special, exemplary, or consequential damages arising from the use of this software.