Amazon Nova Canvas MCP

How to Install and Use Amazon Nova Canvas MCP

If you want to create amazing images using text or colors, Amazon Nova Canvas MCP is a great tool for you. This guide will show you how to install it and start using it step by step.

Step 1: Prepare Your Computer

Before installing, make sure you have some important things ready. You need to install a tool called uv, which helps run the MCP server. Also, you must have Python 3.10 on your computer. Lastly, set up your AWS credentials so the tool can safely connect to Amazon services like Bedrock and Nova Canvas.

Here’s what to do:

• Install uv from Astral’s official site or follow their GitHub install guide.
• Install Python by running this command in your terminal:

  uv python install 3.10
  

• Set your AWS credentials with:

  aws configure
  

  Be sure your AWS account has access to Amazon Bedrock and Nova Canvas.

Step 2: Install Amazon Nova Canvas MCP Server

Now you are ready to install the Amazon Nova Canvas MCP server. This server turns your text prompts into beautiful images. You can install it easily using the uvx command.

Simply run this command in your terminal:

uvx awslabs.nova-canvas-mcp-server@latest

This installs the latest version of the server and allows you to use it right away.

Step 3: Configure the MCP Server

After installation, you need to tell your MCP client how to find and use the server. This configuration is done in a file named mcp.json. For example, if you are using the Amazon Q Developer CLI, open the file here: ~/.aws/amazonq/mcp.json.

Add this snippet inside the "mcpServers" section:

{
  "awslabs.nova-canvas-mcp-server": {
    "command": "uvx",
    "args": ["awslabs.nova-canvas-mcp-server@latest"],
    "env": {
      "AWS_PROFILE": "your-aws-profile",
      "AWS_REGION": "us-east-1",
      "FASTMCP_LOG_LEVEL": "ERROR"
    },
    "disabled": false,
    "autoApprove": []
  }
}

Replace "your-aws-profile" with the name of your AWS profile.

Step 4: Windows Users Installation

If you use Windows, the setup is slightly different because of how commands work. Use this configuration instead:

{
  "mcpServers": {
    "awslabs.nova-canvas-mcp-server": {
      "disabled": false,
      "timeout": 60,
      "type": "stdio",
      "command": "uv",
      "args": [
        "tool",
        "run",
        "--from",
        "awslabs.nova-canvas-mcp-server@latest",
        "awslabs.nova-canvas-mcp-server.exe"
      ],
      "env": {
        "FASTMCP_LOG_LEVEL": "ERROR",
        "AWS_PROFILE": "your-aws-profile",
        "AWS_REGION": "us-east-1"
      }
    }
  }
}

Step 5: Use with Docker (Optional)

You can also run the MCP server inside Docker. After building the image with:

docker build -t awslabs/nova-canvas-mcp-server .

Create a .env file with your AWS credentials like this:

AWS_ACCESS_KEY_ID=YOUR_ACCESS_KEY
AWS_SECRET_ACCESS_KEY=YOUR_SECRET_KEY
AWS_SESSION_TOKEN=YOUR_SESSION_TOKEN

Then configure your MCP client to run the server with this docker command:

{
  "mcpServers": {
    "awslabs.nova-canvas-mcp-server": {
      "command": "docker",
      "args": [
        "run",
        "--rm",
        "--interactive",
        "--env",
        "AWS_REGION=us-east-1",
        "--env",
        "FASTMCP_LOG_LEVEL=ERROR",
        "--env-file",
        "/full/path/to/.env",
        "awslabs/nova-canvas-mcp-server:latest"
      ],
      "env": {},
      "disabled": false,
      "autoApprove": []
    }
  }
}

Make sure to replace /full/path/to/.env with where your file is saved.

Step 6: AWS Authentication

Amazon Nova Canvas MCP uses your AWS profile to securely connect to services. You can set the profile name and region in environment variables like this:

"env": {
  "AWS_PROFILE": "your-aws-profile",
  "AWS_REGION": "us-east-1"
}

This makes sure the MCP server can authenticate using your AWS credentials without exposing your keys elsewhere.

---

Now you’re ready to generate images using the Amazon Nova Canvas MCP server! Just follow these steps, and you can start turning text prompts or colors into fantastic visuals easily.