If you're reading this article, chances are you've already had to **inject a dynamic Tailwind CSS class into an HTML element**, something like this: ```html ``` During the development phase, everything works fine, but as soon as it comes to the "build" phase to put the code into production, **you end up with several Tailwind classes missing!** > But why? To avoid having CSS files filled with unused classes (Tailwind has 1500+ of them), the build phase will detect all the classes actually present in the code, **and generate the final file with only the classes used**. But the problem is that Tailwind is **unable to detect dynamically generated classes** like this `'bg-'+alert.color+'-600'` ## The solutions There are two main solutions, to be chosen depending on your use case. > Don't hesitate to take a look at [the official documentation](https://tailwindcss.com/docs/content-configuration), in the "Content Configuration" chapter, for more details! ### Use full class names The difference may seem trivial in terms of code, and yet it makes all the difference during the build phase. Instead of "generating" class names dynamically, **you should favor using the full names of the classes, and "use" them dynamically**, like this: ```html ``` > Note that class detection **also works in JS/TS scripts** if you use the full names of the classes! ### Configure a safelist In the `tailwind.config.js` configuration file, it's possible **to specify a list of classes to include in the final file**, even if these classes are not found in the code, for example: ```jsx module.exports = { //... safelist: [ 'text-2xl', 'text-3xl', { pattern: /bg-(red|green)-600/, }, ], } ``` > As you can see, it's even possible to **specify [regular expressions](https://code-garage.fr/blog/how-to-improve-with-regexp) to include several classes at once!** Note that it is still advisable to use **the previous solution, to avoid oversights!**