{
  "id": 12583090,
  "title": "How do you give an AI real design taste?",
  "url": "https://urgent.news/2026/10/07/how-do-you-give-an-ai-real-design-taste",
  "topic": "ai",
  "section": "AI",
  "published": "2026-10-07T08:24:16.000Z",
  "source": {
    "name": "Dev.to",
    "slug": "dev-to",
    "url": "https://dev.to/usman_basheers/how-do-you-give-an-ai-real-design-taste-20m6"
  },
  "original_language": "en",
  "account": "Giving an AI real design taste requires a structured process and tools to guide its output. The SVG Lab Design Harness provides a Design Engine, tools, and checks that work within your existing assistant through the SVG Lab MCP (Model Context Protocol). This setup allows you to inspect the model's work and request revisions through the Design Engine's drawing tools, guides, and checks. You treat this as a different tool environment rather than altering the model's weights.\n\nTo connect the SVG Lab MCP to your assistant, you need to add a custom connector in your assistant's settings. For Claude, you can enable Developer mode and add the connection using the provided address and OAuth sign-in. For ChatGPT, you can create an app in Plugins, Apps, or Connectors with the address and OAuth credentials. For other tools like Codex or Cursor, you can follow similar steps to add a custom Streamable HTTP MCP.\n\nOnce connected, you can receive a design brief from your assistant, such as creating a mobile app UI for a creative consumer app with multiple screens, polished modern layout, clear hierarchy, and a strong visual identity. The AI will use the Design Engine with SVG Lab's drawing tools, guides, and checks to create the design. In the example, Claude Opus 5.5 (Anthropic) produced a four-screen app called \"murmur\" featuring vector shapes of a heron, cat, birds, and figures. The entire process took about 40 minutes and 119 design steps, with no edits made after the run.\n\nThe connected workflow supplies a design workspace and process, allowing you to utilize the assistant for finished screen artwork and drawn illustrations. The harness changes the available tools, guidance, and checks, but the underlying model weights remain unchanged. The example demonstrates one connected historical run, not a matched benchmark with the same model without the harness. After receiving the design, you can export the artwork as SVG, PNG, or PDF and implement it in your application code separately.",
  "summary": "An AI needs a design process it can use and inspect: a product brief, visual guidance, drawing tools and a review loop. SVG Lab (svglab.app)'s Design Harness provides its Design Engine, tools and checks to your existing assistant through the SVG Lab MCP. Treat that as a different tool environment, rather than a change to the model's weights. You still choose the product's purpose and inspect what…",
  "key_points": [
    "SVG Lab Design Harness provides Design Engine, tools, and checks",
    "Connect SVG Lab MCP to assistant via custom connector settings",
    "AI creates design using Design Engine with SVG Lab tools"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}