Build a type scale with AI
Choose a base size and a ratio, then generate separate body and heading tokens for CSS or Tailwind. Preview them in the actual typefaces you use.
When you build a type scale with AI, try it first on one real article. Its headline, paragraphs, captions and navigation labels have different jobs. The learn2kern skill takes a base size and a chosen ratio and works out related sizes. It separates body and heading families so a tighter display setting does not accidentally spread through the main text.
Numbers alone cannot tell you whether a particular typeface will read well on a narrow screen. A heading set in a wide family may wrap sooner than another heading at precisely the same size. Generate the tokens, then test them with your actual content before changing a whole site.
What it does
- Compare a gentle scale with a more pronounced hierarchy.
- Set line spacing appropriate to long body copy.
- Give heading and body tokens distinct names.
- Create CSS custom properties or Tailwind font-size values.
- Review the smallest label and largest heading together.
Install and use
- Supply the base size, preferred ratio and roles the interface needs. Use typefaces already available and licensed for your project.
- Ask for a practical comparison, such as a news headline that needs to stand out without wrapping to three short lines on a phone.
- Apply the tokens to a sample page, review real copy at several widths and adjust the scale before wider use.
FAQ
Source
Made by Information Logistics. Original skill: learn2kern, published on GitHub under the MIT licence.
View on GitHub