AWS Labs Nova Canvas MCP Server

1.0.15 verified Fri May 01 auth: no python

An AWS Labs Model Context Protocol (MCP) server for Amazon Nova Canvas, enabling AI agents to generate and edit images via the Nova Canvas model. Current version: 1.0.15. Release cadence: irregular.

pip install awslabs-nova-canvas-mcp-server

Common errors

error botocore.exceptions.NoCredentialsError: Unable to locate credentials ↓

cause AWS credentials not found in environment or passed to client.

fix

Set AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY environment variables or pass them to the NovaCanvasMCPClient constructor.

error ModuleNotFoundError: No module named 'nova_canvas_mcp_server' ↓

cause The package is not installed or imported incorrectly.

fix

Run 'pip install awslabs-nova-canvas-mcp-server' and use 'from nova_canvas_mcp_server import NovaCanvasMCPClient'.

Warnings

gotcha Requires Python >=3.10. Using older Python will cause installation failures. ↓

fix Upgrade Python to 3.10 or later.

deprecated AWS credentials must be provided via environment variables or passed explicitly. The library does not support IAM roles by default. ↓

fix Set AWS_ACCESS_KEY_ID, AWS_SECRET_ACCESS_KEY, and AWS_REGION environment variables.

Imports

NovaCanvasMCPClient

from nova_canvas_mcp_server import NovaCanvasMCPClient

Quickstart

Initialize the MCP server client with AWS credentials and generate an image.

import os
from nova_canvas_mcp_server import NovaCanvasMCPClient

client = NovaCanvasMCPClient(aws_access_key_id=os.environ.get('AWS_ACCESS_KEY_ID'), aws_secret_access_key=os.environ.get('AWS_SECRET_ACCESS_KEY'), region=os.environ.get('AWS_REGION', 'us-east-1'))
result = client.generate_image(prompt='A sunset over mountains')
print(result)