Workers MCP Image Generation Service & Hello World

Overview

This project implements a Cloudflare Worker that uses the Model Completion Protocol (MCP) to provide a simple API for image generation. The worker exposes two main endpoints:

• sayHello: A simple greeting function
• generateImage: An endpoint that leverages Cloudflare's AI capabilities to generate images using the flux-1-schnell model

Features

• Text-to-image generation using Flux model
• Configurable diffusion steps for quality control
• Simple API interface
• Built on Cloudflare Workers platform

Requirements

• Cloudflare Workers account
• AI binding configuration in your Cloudflare environment

Installation

1. Clone this repository
2. Install dependencies:

pnpm install

3. Configure your wrangler.toml file with the necessary AI bindings:

[[ai_bindings]]
name = "AI"

Usage

Once deployed, you can interact with the API using the following endpoints:

Say Hello

POST /sayHello
{
  "name": "Your Name"
}

Generate Image

POST /generateImage
{
  "prompt": "A detailed description of the image you want to generate",
  "steps": 30
}

The steps parameter controls the number of diffusion steps. Higher values (20-50) will produce better quality images but take longer to generate.

Code Structure

• Uses the Workers MCP framework for handling API requests
• Implements TypeScript interfaces for environment variables
• Leverages Cloudflare's AI platform for image generation
• Converts base64 image data to binary for proper image serving

Error Handling

The code includes error handling for:

• Missing AI binding configuration
• Failed image generation requests
• Other runtime errors

Development

To run this project locally:

pnpm run dev

To deploy to production:

pnpm run deploy