Midjourney MCP Server

How to use

Connect to your MCP server from your client of choice by configuring the MCP server entry with the command to start the local server and the required environment variables. You can run the server locally for testing or production use and then issue commands to generate images, edit images, translate prompts, and create videos using the available tools.

How to install

# Prerequisites
- Python 3.9+ and pip
- Or use the uv transport for a smoother development experience

# Clone the repository
git clone https://github.com/AceDataCloud/mcp-midjourney.git
cd mcp-midjourney

# Install with pip (editable mode)
pip install -e .

# Or install with uv transport (recommended)
uv pip install -e .

Claude Desktop Integration

Add a MCP server configuration to Claude Desktop to enable one-click access from Claude. The configuration supports both a direct MCP command and a uv-based run command.

{
  "mcpServers": {
    "midjourney": {
      "command": "mcp-midjourney",
      "env": {
        "ACEDATACLOUD_API_TOKEN": "your_api_token_here"
      }
    }
  }
}

Environment and startup options

The server relies on environment variables for API access and configuration. You can run the server in a standard stdio mode or using uv as the transport. Use the provided tokens and optional base URL to point to the AceDataCloud API.

# Example environment variables for a local run
ACEDATACLOUD_API_TOKEN=your_api_token_here
ACEDATACLOUD_API_BASE_URL=https://api.acedata.cloud
MIDJOURNEY_DEFAULT_MODE=fast
MIDJOURNEY_REQUEST_TIMEOUT=180
LOG_LEVEL=INFO

Available tools and usage patterns

The server exposes a set of tools to perform image and video tasks, translations, and task queries. Use the tools to generate images from prompts, transform and blend images, describe or edit images, generate or extend videos, translate prompts, and monitor task progress.

Configuration

Configure the server with the required API token and optional defaults to tailor generation behavior.

# Example: environment variables for runtime
ACEDATACLOUD_API_TOKEN=your_api_token_here
ACEDATACLOUD_API_BASE_URL=https://api.acedata.cloud
MIDJOURNEY_DEFAULT_MODE=fast
MIDJOURNEY_REQUEST_TIMEOUT=180
LOG_LEVEL=INFO

Notes on generation modes

Choose a mode based on your needs. fast is the default, turbo accelerates generation with higher resource use, and relax provides a cheaper, slower option.

Troubleshooting tips

If you encounter connectivity or token issues, verify your ACEDATACLOUD_API_TOKEN and ensure your base URL is reachable. Check that the server process is running and listening on the configured transport, port, and path.

Tools snapshot

The MCP server provides a suite of tools grouped by function, including image generation, image editing, video creation, translation, and task management.

midjourney_imagine

Generate images from a text prompt and create a 2x2 grid of results

midjourney_transform

Transform images including upscale, variations, zoom, and pan actions

midjourney_blend

Blend multiple images together into a single output

midjourney_with_reference

Generate using a reference image as inspiration

midjourney_edit

Edit an existing image with a text prompt or masks

midjourney_describe

Describe an image to create a reverse prompt

midjourney_generate_video

Generate a video from text prompts and reference images

midjourney_extend_video

Extend an existing video to make it longer

midjourney_translate

Translate Chinese prompts to English for generation

midjourney_get_task

Query a single task status

midjourney_get_tasks_batch

Query multiple tasks at once

midjourney_list_actions

List available API actions

midjourney_get_prompt_guide

Get prompt writing guidance

midjourney_list_transform_actions

List transformation actions