AI Code Agent

A powerful CLI tool for AI-powered code generation and file management. Similar to Claude Code and Codex, this tool allows you to interact with multiple AI providers to generate code and automatically create files.

Features

• Multi-Provider Support: OpenAI, Anthropic Claude, DeepSeek, Grok (xAI), Kimi (Moonshot)
• Interactive chat interface with AI
• Automatic file creation and code writing
• Persistent conversation history
• Global configuration management
• Easy provider switching
• Token/API key management

Installation

From npm (once published)

npm install -g ai-code-agent

Local Development

# Clone the repository
git clone <your-repo-url>
cd code-agents

# Install dependencies
npm install

# Link the package globally
npm link

Quick Start

1. Configure an AI Provider

Interactive setup (recommended):

ai-agent config set

Direct setup:

# OpenAI
ai-agent config set-token openai sk-your-api-key

# Anthropic Claude
ai-agent config set-token anthropic sk-ant-your-api-key

# DeepSeek
ai-agent config set-token deepseek your-api-key

# Grok (xAI)
ai-agent config set-token grok your-api-key

# Kimi (Moonshot)
ai-agent config set-token kimi your-api-key

2. Set Default Provider (Optional)

ai-agent config default openai

3. Start Chatting

ai-agent

Configuration Commands

View All Supported Providers

ai-agent config providers

List Configured Providers

ai-agent config list

Set API Key with Custom Model

ai-agent config set-token openai sk-your-key --model gpt-4
ai-agent config set-token anthropic sk-ant-key --model claude-3-opus-20240229

Remove Provider Configuration

ai-agent config remove openai

Use Specific Provider for a Session

ai-agent chat --provider anthropic
ai-agent --provider deepseek

Usage Examples

Start Interactive Chat

ai-agent

Creating a file:

You: Create a React component for a login form
AI: I'll create a React component for you...
Create file "LoginForm.jsx"? Yes
✓ File created: LoginForm.jsx

General questions:

You: How do I use async/await in JavaScript?
AI: Async/await is a modern way to handle asynchronous operations...

Exit:

You: exit

Switch Providers Mid-Session

# Use Anthropic Claude
ai-agent --provider anthropic

# Use DeepSeek
ai-agent --provider deepseek

Supported AI Providers

| Provider | ID | Models | |----------|----|--------------------| | OpenAI | openai | gpt-4-turbo-preview, gpt-4, gpt-3.5-turbo | | Anthropic Claude | anthropic | claude-3-5-sonnet-20241022, claude-3-opus-20240229, claude-3-sonnet-20240229 | | DeepSeek | deepseek | deepseek-chat, deepseek-coder | | Grok (xAI) | grok | grok-beta | | Kimi (Moonshot) | kimi | moonshot-v1-8k, moonshot-v1-32k, moonshot-v1-128k |

Project Structure

code-agents/
├── bin/
│   └── cli.js                    # CLI entry point
├── src/
│   ├── api/
│   │   ├── providers/
│   │   │   ├── openai.js        # OpenAI provider
│   │   │   ├── anthropic.js     # Anthropic provider
│   │   │   ├── deepseek.js      # DeepSeek provider
│   │   │   ├── grok.js          # Grok provider
│   │   │   └── kimi.js          # Kimi provider
│   │   └── providerFactory.js   # Provider factory
│   ├── commands/
│   │   └── config.js            # Config commands
│   ├── utils/
│   │   ├── config.js            # Configuration management
│   │   └── fileManager.js       # File operations
│   └── index.js                 # Main application logic
├── package.json
└── README.md

Configuration

The tool stores configuration in ~/.ai-code-agent/config.json:

{
  "defaultProvider": "openai",
  "providers": {
    "openai": {
      "apiKey": "sk-your-openai-key",
      "model": "gpt-4-turbo-preview"
    },
    "anthropic": {
      "apiKey": "sk-ant-your-anthropic-key",
      "model": null
    },
    "deepseek": {
      "apiKey": "your-deepseek-key",
      "model": "deepseek-coder"
    }
  },
  "maxTokens": 2000,
  "temperature": 0.7
}

Requirements

• Node.js 18 or higher
• API key for at least one supported AI provider

Development

# Install dependencies
npm install

# Run locally
npm start

# Test the CLI
node bin/cli.js

How It Works

1. Interactive Prompt: The tool opens an input prompt when you run the command
2. AI Processing: Your prompt is sent to your configured AI provider
3. Smart Response: The AI analyzes your request and determines if it needs to create a file
4. File Creation: If code generation is needed, you'll be prompted to confirm file creation
5. Automatic Writing: The code is automatically written to the specified file

Getting API Keys

• OpenAI: https://platform.openai.com/api-keys
• Anthropic: https://console.anthropic.com/
• DeepSeek: https://platform.deepseek.com/
• Grok (xAI): https://x.ai/api
• Kimi (Moonshot): https://platform.moonshot.cn/

Publishing to npm

# Login to npm
npm login

# Publish the package
npm publish

License

MIT

Contributing

Contributions are welcome! Please feel free to submit a Pull Request.