Amazon Bedrock

You run an MCP server that talks to Amazon Bedrock’s Nova Canvas model to generate images from text. It lets you fine‑tune outputs with negative prompts, control size and quality, and reproduce results with a seed. This guide shows you how to use the server with an MCP client, install and run it locally, and configure credentials and integration points.

How to use

With the Amazon Bedrock MCP Server you can generate images by sending organized prompt data to the MCP endpoint. Use a client to request image generations by describing the scene you want and, if needed, refine composition with a negativePrompt to exclude undesired elements. You can specify width, height, and quality to suit your use case, and set a seed for deterministic results. Request multiple images in a single call when you want a batch from one prompt.

How to install

Prerequisites you need before installing the server.

• Active AWS account with Bedrock and Nova Canvas access
• AWS credentials configured with required permissions
• Node.js version 18 or later

Install and run the MCP server locally with the following steps.

# Clone the MCP server repository
git clone https://github.com/zxkane/mcp-server-amazon-bedrock.git
cd mcp-server-amazon-bedrock

# Install dependencies
npm install

# Build the project
npm run build

# Start the server (example, adapt to your setup if needed)
npx @zxkane/mcp-server-amazon-bedrock

Additional notes and configuration

AWS credentials can be supplied in multiple ways. You can export environment variables, use a credentials file, or rely on an IAM role when running on AWS infrastructure.

# Method 1: Environment variables
export AWS_ACCESS_KEY_ID=your_access_key
export AWS_SECRET_ACCESS_KEY=your_secret_key
export AWS_REGION=us-east-1

# Method 2: Credentials file (default profile)
# ~/.aws/credentials
[the_profile_name]
aws_access_key_id = your_access_key
aws_secret_access_key = your_secret_key
export AWS_PROFILE=the_profile_name

Integration with Claude Desktop

If you want Claude Desktop to work with this MCP server, add the following configuration to your settings. This example uses a local stdio MCP server via npx and passes the necessary AWS variables.

{
  "mcpServers": {
    "amazon-bedrock": {
      "command": "npx",
      "args": [
        "-y",
        "@zxkane/mcp-server-amazon-bedrock"
      ],
      "env": {
        "AWS_PROFILE": "your_profile_name",
        "AWS_ACCESS_KEY_ID": "your_access_key",
        "AWS_SECRET_ACCESS_KEY": "your_secret_key",
        "AWS_REGION": "us-east-1"
      }
    }
  }
}

Security and best practices

Keep your AWS credentials secure. Prefer IAM roles when running on AWS infrastructure and avoid embedding secrets in code or configuration files you share publicly. Use the minimum required Bedrock permissions for your tasks and rotate keys regularly.

Available tools

generate_image

Creates images from text descriptions using Amazon Bedrock's Nova Canvas model. Accepts prompt, negativePrompt, width, height, quality, cfg_scale, seed, and numberOfImages to produce one or more images with configurable parameters.