AI Integration

dartwork-mpl is built for the age of AI-assisted coding. Every API, every default, and every bundled asset is designed so that AI coding assistants produce correct, publication-quality plots on the first try.

How It Works

dartwork-mpl provides three layers of AI integration, each building on the last:

Layer 3 — MCP Server 🔌

AI reads docs in real time via Model Context Protocol

Layer 2 — Bundled Guides 📖

get_prompt · copy_prompt · install_llm_txt

Layer 1 — AI-Friendly API 🎯

dm.style.use · dm.simple_layout · oc.blue5

Each layer builds on the one below.

Layer 1 works out of the box — no setup required. AI assistants naturally produce better code because the API has fewer ways to go wrong.

Layer 2 adds bundled documentation that AI can read programmatically, even without internet access.

Layer 3 provides real-time documentation access through the Model Context Protocol, so the AI always has the latest guidelines.

Quick Start

Get up and running with AI-assisted plotting in 30 seconds:

Add the dartwork-mpl MCP server to your AI client config, and it will automatically access documentation and guidelines:

{
  "mcpServers": {
    "dartwork-mpl": {
      "command": "uv",
      "args": [
        "run",
        "--directory",
        "/path/to/dartwork-mpl",
        "dartwork-mpl-mcp"
      ]
    }
  }
}

Then just ask your AI assistant to create a plot — it will know exactly how to use dartwork-mpl.

→ Full setup: MCP Server

Install static guide files into your IDE’s AI context folders:

import dartwork_mpl as dm
dm.install_llm_txt()
# ✅ Guides installed for Claude Code and Cursor IDE

→ Details: AI-Assisted Development

Even without any configuration, dartwork-mpl’s consistent API makes AI-generated code more reliable. Just install the package and tell your AI assistant to use it:

import dartwork_mpl as dm

fig, ax = dm.subplots(style=['font-scientific'])
ax.plot(x, y, color='oc.blue5')
dm.simple_layout(fig)
dm.save_and_show(fig)

→ Learn why: Why AI-Ready?

What’s Inside