Paid prompts or free components: which one you need

A component library gives you a component. A design prompt gives you the decisions and no code. Those are different objects solving different halves of the same job, and buying one while needing the other is the most common way to lose a weekend on this.

A component library hands you a card that flips, a marquee that scrolls, a gradient border that animates. You install it, you import it, it works. Free, and for a great many projects exactly right.

A prompt hands you none of that. It names the ground, the surface, the ink and the accent. It names two faces with the fallback stack written out and a scale in px, a curve with a real value and a duration in milliseconds, and the order the sections run in. Then a model writes the code, and the code is yours. What arrives, side by side:

The corollary is blunt and it settles most of the argument. If your page is short of components, a prompt will not help you. If your page is short of decisions, a component will not help you. Most pages that look generated are short of decisions.

The free layer is bigger than the paid one#

Worth saying plainly on a page that ends up recommending something paid.

The free side of this is not a scrappy alternative. It is the larger half of the market:

• Component libraries, genuinely good, mostly MIT, at zero.
• Free prompt libraries publishing a hundred or more complete prompts each with no account and no email.
• The builders' own documentation, which runs to thousands of words of prompting and design guidance and costs nothing, and of which Lovable's is the largest.
• The general prompt repositories on GitHub, carrying six figures of stars between them.

In the corner of this market where both sides are visible, free downloads outnumber paid purchases by more than twenty to one, and the median price anybody actually pays is around $19.

That ratio is the honest backdrop to any paid recommendation, including this one. Anything sold here is competing against a large, functional, free alternative, not against nothing.

And the components being free is not a weakness of theirs to be worked around. It is the correct price for what they are. A component library is built to be reusable, an opinion about your palette would make it less reusable, and reusability is the entire product.

Where a free component stops#

A component arrives with a default, and defaults are the measured cause of the sameness everybody complains about.

• Anthropic's engineering blog says its model, unguided, reaches for "Inter fonts, purple gradients on white backgrounds, and minimal animations".
• Bolt's blog says AI will almost always pick Inter or Roboto and safe blue and grey, and that this is why a lot of AI designs blend together.
• v0's system prompt bans purple and gradients by name, and mentions animation, easing, duration and reduced motion exactly zero times.
• Across all the design writing the major builders publish, there is one real duration table and no published easing curve values at all.

Installing a beautiful free component gives you a beautiful default, and a beautiful default installed by ten thousand people is still a default. That is not the library's fault and it is not something the library is trying to solve. Deciding your palette is not its job.

One caution for the buyer, at the level of the whole category. Some paid prompt catalogues include repackaged components that are MIT licensed and already free. Before paying anybody, open a complete item and ask whether what you are being sold is direction or a rewrapped package you could have installed yourself in a minute. That question is worth asking of every one of the libraries selling these.

When free is enough, and it often is#

This is a decision, not a hedge. Free is the right answer when:

• You already have a design system. The decisions are made. You need parts, and a component library plugs straight into tokens you have decided.
• You are prototyping to answer a question that is not about appearance. Nobody is going to judge the type scale of a thing you will throw away on Friday.
• The page is one screen and one call to action. There is nothing to differentiate, so the default is fine and faster.
• Nobody will compare your page to a competitor's. Internal tools, admin panels, anything with a captive audience.

Paid direction earns its price in one situation: when the page has to survive being put next to a competitor's, and the person paying can see that one page reads as expensive and the other does not but cannot say why. That is a narrow claim, and it is the only one worth making.

What to check before you pay for either#

Five minutes, and it applies to both classes.

1. Open the page source. If the paid text is already sitting in it behind a copy button, you are buying a button.
2. Count the free tier. How many complete artefacts can you take without an account? In this category the answer is often a lot.
3. Look for a number. A named curve with its value, a duration in milliseconds, a hex with a role. "Smooth animations" is not a decision.
4. Check the licence. MIT components cost nothing. Ask precisely what is being added on top of them.
5. Check the refund. It ranges from fourteen days no questions to "All sales are final". Both are in writing before you pay.

Which one you need#

• Short of parts? Take the free component libraries. They are good, they are MIT, and paying for an alternative would be a waste.
• Short of decisions? No component fixes that, and neither does a longer adjective. You need the values written down: the palette with roles, the scale, the curve, the duration, the order, and the list of what this design refuses to do.

That second one is the product Thanor sells: full art direction per design, in exact values, for UI, app and SaaS interfaces, page sections and animated backgrounds, usable on personal and client work. $89 for three months, $189 a year, $299 once.

Free accounts open the designs marked free, so you can run every check on the list above against us first. Start there.