Now available

The drawing engine built for AI

OmniDraw lets your AI assistant create, edit, and export images — right from the chat. Install it once, and start drawing with natural language.

One-time setup

Up and running in under a minute

Install with pip, add one line to your MCP client config, and you're done. OmniDraw shows up as a drawing tool your AI can call instantly.

Quick start

Two commands and one config — that's it.

# 1. Install
pip install omnidraw

# 2. Add to your MCP client config
{
  "mcpServers": {
    "omnidraw": {
      "command": "omnidraw"
    }
  }
}

# 3. Reload your MCP client — done!

Why OmniDraw

Purpose-built for AI-driven image creation. No GUI needed — your AI does the drawing while you focus on the idea.

AI draws, you get images

Tell your AI assistant what to draw. OmniDraw handles shapes, text, colors, and layout — then exports pixel-perfect images.

Organize with projects

Keep your work structured. One project holds multiple compositions and shared assets so nothing gets lost.

Reusable assets

Import images once, reuse them across compositions. No need to re-upload the same logo or photo every time.

Runs 100% locally

Your data stays on your machine. OmniDraw runs as a local server — no cloud uploads, no API keys, no latency.

Works with any MCP client

OmniDraw speaks the Model Context Protocol. Use it with Cursor, Claude Desktop, or any MCP-compatible AI client.

From idea to export in seconds

Create a project, describe what you want, and export production-ready images — all within one conversation.

Free to start, Pro when ready

The free tier covers the core drawing workflow. Upgrade to Pro for unlimited projects, asset management, and advanced exports.

What's next

On the roadmap

OmniDraw is actively evolving. Here's what's coming soon.

More drawing primitives

Gradients, shadows, curves, clipping masks, and richer text styling for more expressive compositions.

Templates and smart layouts

Pre-built templates for social media posts, banners, diagrams, and more — so your AI can start with a solid foundation.