Image_Generation

🎨 Serverless Image Generator (Bedrock + Stable Diffusion)

📖 Overview

This project implements a fully Serverless GenAI Architecture. Instead of running heavy GPU instances to host image models, this solution leverages Amazon Bedrock to generate images via API calls. It exposes a REST API that accepts text prompts and returns secure, pre-signed S3 URLs for the generated images.

🚀 Key Features

• Zero Infrastructure Management: purely serverless (Lambda + API Gateway).
• Secure Delivery: Images are stored in a private S3 bucket and accessed only via time-limited Pre-signed URLs.
• Scalable: Handles concurrent requests automatically via AWS Lambda.

🏗️ Architecture

⚙️ How it Works

1. User sends a POST request to API Gateway: {"prompt": "A futuristic city"}.
2. Lambda triggers Bedrock (Stable Diffusion model).
3. Bedrock returns base64 image data.
4. Lambda decodes and saves the image to a private S3 Bucket.
5. Lambda generates a Pre-signed URL (valid for 5 mins) and returns it to the user.

💻 Usage (Postman)

Endpoint: POST https://<your-api-id>.execute-api.us-east-1.amazonaws.com/prod/generate

Body:

{
  "prompt": "A cyberpunk workspace with neon lights, 8k resolution"
}

Response:

{
  "status": "success",
  "image_url": "https://my-bucket.s3.amazonaws.com/gen-123.jpg?AWSAccessKeyId=..."
}

🧠 Key Learnings & Patterns

• Handling Binary Data: API Gateway has a 10MB payload limit. Passing base64 images directly through the API response is bad practice. The "S3 Pre-signed URL" pattern used here is the enterprise-standard way to deliver large assets securely and efficiently.
• Cold Starts: I optimized the Lambda function by moving the boto3 client initialization outside the handler to reuse connections across warm invocations.

---

Maintained by Phani Kolla

---