🧠 AI Code Generator using AWS Bedrock + Claude

This is a simple yet powerful Python-based AI tool that generates code from text instructions using Anthropic Claude (via AWS Bedrock) and automatically saves the output to an Amazon S3 bucket.

Ideal for serverless deployment on AWS Lambda, but also adaptable for local use.

---

✨ Features

• Uses Claude v2 on AWS Bedrock for natural code generation
• Input prompt: "Write Python code for reversing a list"
• Output: Saved as .py file in S3
• Fully serverless when deployed on Lambda
• Lightweight, fast, and scalable

---

📁 Files & Their Roles

File	Purpose
code_generation.py	Main script: generates code and saves it to S3
boto3Layer.script	Bash script: creates a boto3 Lambda layer
boto3_layer.zip	Pre-built layer to upload to AWS Lambda
README.md	Full setup and usage documentation
requirements.txt	Optional: Install locally for testing
event_sample.json	Sample event input for local testing

---

🚀 How It Works

1. User sends a text prompt (e.g. “Build a login form in HTML”)
2. Claude AI processes the prompt and returns code
3. The code is stored as a .py file in your specified S3 bucket

---

⚙️ Setup Instructions (AWS Lambda)

✅ Prerequisites

• An AWS account with access to:
  • Bedrock
  • S3
  • Lambda
• An existing S3 bucket (or create one)
• Python 3.x environment (for local testing)

---

🛠️ Step 1: Create the Boto3 Lambda Layer

Run this command in your terminal (on Mac/Linux):

bash boto3Layer.script

This will:

• Create a virtual environment
• Install boto3 into a python/ directory
• Zip it into boto3_layer.zip

---

🧩 Step 2: Upload Layer to AWS Lambda

1. Go to AWS Console → Lambda → Layers → Create layer
2. Name it boto3-layer
3. Upload boto3_layer.zip
4. Select your Python runtime (e.g., Python 3.9)
5. After creating, go to your Lambda function → Add layer → select the one you just created

---

📜 Step 3: Deploy the Lambda Function

1. Create a new Lambda function
2. Paste the contents of code_generation.py into the Lambda code editor
3. Set the handler to:

lambda_handler

4. Set Python runtime (same as your layer, e.g. Python 3.9)
5. Attach your boto3 layer to the function

---

🧠 Step 4: Configure Bedrock + S3

Inside code_generation.py, edit the following:

s3_bucket = 'your-bucket-name'  # Replace with your real bucket

You can also change:

• AWS region: us-west-2 → your region
• Model ID: e.g., anthropic.claude-v2 → another Bedrock model

---

🧪 Testing the Function

Use the Lambda test tab or an API Gateway to send this payload:

{
  "body": {
    "message": "Write a Python class for a calculator",
    "key": "Python"
  }
}

✅ Expected Result:

A new file appears in your S3 bucket under code-output/HHMMSS.py, containing Claude’s generated code.

---

🔁 Run Locally (Optional)

You can adapt the script for local use. Just simulate the Lambda input like this:

event = {
    "body": json.dumps({
        "message": "Create a web server in Node.js",
        "key": "JavaScript"
    })
}

lambda_handler(event, None)

Make sure your local machine has:

pip install boto3
aws configure  # Add your credentials

---

📦 Project Improvements

• Add HTML/React front-end to collect prompt + display code
• Add support for multiple AI models (Claude, Jurassic, Titan, etc.)
• Return generated code via API (instead of only saving to S3)

---

📝 License

MIT License — Free to use and modify.