Amazon Nova Reel

You can run an MCP server for Amazon Nova Reel 1.1 to generate video content asynchronously, monitor progress, and retrieve final outputs. It supports multiple transport methods (stdio, SSE, and HTTP streaming), integrates with AWS Bedrock and S3, and includes a prompting guide to help you craft effective prompts for video generation.

How to use

You interact with the Nova Reel MCP server through an MCP client to start new video generation jobs, check their status, and fetch the resulting video URLs. Start a job by providing a textual prompt and optional parameters such as duration, frames per second, and video dimensions. Track all active and completed jobs, then retrieve the video URL once a job finishes. You can choose among transports: stdio for local process communication, SSE for a live streaming interface, or HTTP streaming for bidirectional transport.

How to install

Prerequisites you need to meet before installation:

• Python 3.8+

• AWS Bedrock access and an S3 bucket for video output

• AWS credentials with appropriate permissions

Steps for Docker-based setup (recommended)

# STDIO version
# Run the STDIO transport container
docker run -i mirecekd/novareel-mcp-server:stdio \
  --aws-access-key-id YOUR_AWS_ACCESS_KEY_ID \
  --aws-secret-access-key YOUR_AWS_SECRET_ACCESS_KEY \
  --s3-bucket YOUR_S3_BUCKET_NAME

# SSE version
# Run the SSE transport container
docker run -p 8000:8000 \
  -e AWS_ACCESS_KEY_ID=YOUR_KEY \
  -e AWS_SECRET_ACCESS_KEY=YOUR_SECRET \
  -e S3_BUCKET=YOUR_BUCKET \
  mirecekd/novareel-mcp-server:sse

# HTTP Streaming version
# Run the HTTP transport container
docker run -p 8001:8001 \
  -e AWS_ACCESS_KEY_ID=YOUR_KEY \
  -e AWS_SECRET_ACCESS_KEY=YOUR_SECRET \
  -e S3_BUCKET=YOUR_BUCKET \
  ghcr.io/mirecekd/novareel-mcp:latest-http

Note: Replace the placeholders with your actual AWS credentials and bucket name.

## Configuration and environment

Set the following environment variables to enable AWS Bedrock access and S3 output:

AWS_ACCESS_KEY_ID=YOUR_AWS_ACCESS_KEY_ID AWS_SECRET_ACCESS_KEY=YOUR_AWS_SECRET_ACCESS_KEY S3_BUCKET=YOUR_S3_BUCKET_NAME

## Usage with an MCP client

Once the server is running, use your MCP client to start a new video generation job by providing a prompt and optional parameters. You can monitor progress with list and get operations and, when a job completes, obtain the video URL stored in your S3 bucket.

## Prompting guidance

A comprehensive prompting guide helps you craft effective prompts for video generation. Use clear, specific descriptions, camera terminology, lighting details, and structure prompts to fit the desired duration.

## Examples and tips

For best results, match the prompt complexity to the video length and include scene structure, lighting, and camera movement. The guidance is designed to help you generate engaging video prompts that align with AWS Bedrock capabilities.

## Troubleshooting and notes

If you encounter issues, verify your AWS credentials, ensure the S3 bucket exists and is accessible in the correct region, and confirm the selected transport method is running. Check logs for any errors related to Bedrock or S3 access.

## Package and build notes

You can build and run containers locally or via docker-compose. Ensure you provide the correct AWS credentials and S3 bucket in all deployment methods.

## Available tools

### start\_async\_invoke

Starts a new asynchronous video generation job with a text prompt and optional parameters such as duration, frame rate, and dimension. Returns job identifiers for tracking.

### list\_async\_invokes

Lists all tracked video generation jobs and their current statuses, enabling you to monitor progress at a glance.

### get\_async\_invoke

Fetches detailed information about a specific video generation job, including status and the video URL when complete.

### get\_prompting\_guide

Provides a comprehensive prompting guide with best practices, examples, and templates to improve video generation results.