Ai Code Generator

AI code generator using Plan-and-Solve + ReAct for generating complete, runnable code from requirements and specifications.

AI-powered code generation tool that combines structured planning (Plan-and-Solve) with tool use (ReAct) to generate complete, production-ready code from natural language requirements.

---

Features

📝 Requirement Analysis

• Understanding: Parse natural language requirements
• Clarification: Ask clarifying questions when needed
• Specification: Generate technical specification

🏗️ Code Generation

• Full-stack Support: Frontend, backend, database
• Multiple Languages: JavaScript, Python, TypeScript, Go
• Best Practices: Clean code, design patterns
• Complete Projects: Full project structure

🔧 Tool Integration

• File Generation: Create multiple files
• Dependency Management: package.json, requirements.txt
• Testing: Generate unit tests
• Documentation: README, API docs

---

Usage

Basic Code Generation

javascript

const generator = new CodeGenerator();

const project = await generator.generate({
  requirements: 'Create a REST API for a todo app with user authentication',
  language: 'javascript',
  framework: 'express',
  database: 'mongodb'
});

console.log(project.files);
console.log(project.instructions);

Advanced Options

javascript

const generator = new CodeGenerator({
  style: 'professional',
  includeTests: true,
  includeDocs: true,
  verbose: true
});

const project = await generator.generate({
  requirements: 'Build a real-time chat application',
  language: 'typescript',
  framework: 'nestjs',
  database: 'postgresql',
  features: ['websocket', 'jwt-auth', 'message-history']
});

---

Example Output

text

project/
├── src/
│   ├── controllers/
│   │   └── todo.controller.js
│   ├── models/
│   │   └── todo.model.js
│   ├── routes/
│   │   └── todo.routes.js
│   └── middleware/
│       └── auth.middleware.js
├── tests/
│   └── todo.test.js
├── package.json
├── .env.example
└── README.md

---

Workflow

text

User Requirements
    ↓
Plan-and-Solve Agent
    ├─ Phase 1: Analyze requirements
    ├─ Phase 2: Design architecture
    ├─ Phase 3: Plan file structure
    └─ Phase 4: Generate code
    ↓
ReAct Agent (for each file)
    ├─ Research best practices
    ├─ Generate code
    ├─ Review and fix
    └─ Write to file
    ↓
Complete Project

---

Installation

bash

clawhub install ai-code-generator

---

License

MIT

---

Version

1.0.0

---

Created

2026-04-02