mcp -amazon-bedrock

This MCP server provides an integration with Amazon Bedrock’s Nova Canvas model to generate images from text prompts. It exposes an generate_image tool that accepts a prompt, optional negativePrompt to constrain the output, and controls for image dimensions, quality, seed, and batch size. You can invoke this tool to produce single or multiple images per request, enabling deterministic results when a seed is supplied. The server is designed to validate inputs and handle errors gracefully, ensuring reliable image generation workflows for design, concept art, and rapid prototyping.

To use it, configure your MCP client or integrator to call the generate_image tool with the required prompt and any optional parameters. If you need to restrict or refine output, leverage negativePrompt to steer composition and rely on width, height, quality, cfg_scale, and seed to control resolution, fidelity, and reproducibility. The integration also supports batch generation via numberOfImages, which is useful for creating variations or multiple concepts in one request.

Prerequisites:

• An active AWS account with Bedrock access (Nova Canvas) and valid credentials
• Node.js v18 or later installed on your system
• Internet access to fetch the MCP server package from npm

Installation steps:

1. Clone the repository (if you have source access) or install the MCP server package directly:

   • Cloning (example): git clone https://github.com/zxkane/mcp-server-amazon-bedrock.git cd mcp-server-amazon-bedrock

2. Install dependencies: npm install

3. Build the server (if the project provides a build step): npm run build

4. Start the server using npm (default startup script): npm start

Prerequisites recap:

• Ensure AWS credentials are configured (environment variables, credentials file, or IAM role) with access to Amazon Bedrock Nova Canvas
• Have npm and Node.js installed and accessible in your shell

• Ensure AWS credentials are configured prior to starting the server. Reflect the credentials in environment variables, the AWS credentials file, or an IAM role when deployed on AWS infrastructure.
• The generate_image tool supports width, height, quality, cfg_scale, and seed to control output characteristics. Higher resolutions and multiple images may increase generation time and cost.
• If using Claude Desktop integration, you can add the provided json snippet to the Claude config to expose the amazon-bedrock MCP server within Claude.
• Validate that your AWS region matches the Bedrock endpoint region you intend to use (e.g., us-east-1).