Install

Get OmniDraw running in minutes

Install OmniDraw with one command, add it to your MCP client, and start creating images with AI right away.

Step 1 — Install OmniDraw

OmniDraw is distributed as a Python package. Make sure you have Python 3.10+ installed, then run:

pip install omnidraw

This installs the omnidraw command on your system.

Step 2 — Add to your MCP client

Add the following entry to your MCP client's server configuration (e.g. Cursor, Claude Desktop, or any MCP-compatible client):

{
  "mcpServers": {
    "omnidraw": {
      "command": "omnidraw"
    }
  }
}

Reload or restart your MCP client. OmniDraw will appear as an available tool your AI can call.

Step 3 — Start drawing

Once connected, just tell your AI what you want to draw. OmniDraw provides tools for your AI to:

  • Create and manage projects
  • Add compositions with custom sizes
  • Draw shapes, text, and images
  • Inspect canvas state
  • Undo / redo changes
  • Export images as PNG

System requirements

Python

3.10 or later

OS

Windows, macOS, or Linux

MCP client

Any MCP-compatible AI client

Pro license activation

To unlock Pro features (unlimited compositions, asset import, project save/open, supersampling), set your license key as an environment variable:

# Set your Pro license key
export OMNIDRAW_LICENSE_KEY="your-license-key"

# Then restart your MCP client

Don't have a license key yet? View pricing plans