Is AI killing the plugin market?

For years, one of the first questions when building a website and needing a particular feature has been: is there a plugin for that?

Need to add a particular field to the CMS? Find a plugin.

Need to connect to an external service? Find a plugin.

Need a slightly different way of displaying some content? There is probably a plugin for that too.

Plugins and add-ons have been an important part of web development for a long time. They let developers avoid reinventing the wheel and give site owners access to functionality that would otherwise be expensive to build from scratch.

But AI is starting to change the equation, particularly at the smaller end of the plugin market.

Sometimes it's now easier to just build it

There are plenty of little website features that sit in an awkward middle ground.

They're not difficult enough to justify a large development project, but they can still take enough time that, historically, it made sense to look around for an existing plugin.

Maybe you want to add a small CMS utility. Or change the way something is sorted. Or pull some information from an API. Or add a very specific piece of functionality that only applies to one website.

Previously I might have spent half an hour searching for an add-on, reading the documentation, checking whether it was still maintained and then working out whether it could be bent into doing exactly what I wanted.

Increasingly, I'll just build it.

AI has made that significantly easier.

It can help work through an unfamiliar API, scaffold some of the repetitive code, explain the parts of a framework I don't use every day or help me quickly test a few different approaches.

The result is that something which may once have justified buying and installing a plugin can sometimes be written specifically for the project in roughly the same amount of time.

It changes the cost calculation

Plugins have never really been free, even when they don't have a price tag.

There is always some cost involved.

You need to understand how they work. You need to configure them. You need to keep them updated. You need to make sure they're compatible with the rest of your website. And you are relying on someone else continuing to maintain them.

For larger or more complicated functionality, that trade-off still makes complete sense.

I have no interest in building my own payment gateway, search engine or image manipulation library just because AI can help me write code faster.

There is enormous value in mature software that has been tested across thousands of websites, handles edge cases I haven't thought about and is maintained by people who specialise in that particular problem.

But for something small and very specific, the balance is changing.

If I can write 50 lines of code that do exactly what the website needs, with no additional dependencies and no configuration screen full of options I'll never use, that starts to look pretty attractive.

AI hasn't removed the need to know what you're doing

This is probably the most important part.

AI making it easier to write code doesn't necessarily mean everyone should replace plugins with AI-generated code.

You still need to understand what the code is doing.

You need to know whether it is secure. Whether it handles errors properly. Whether it follows the conventions of the platform you're working with. Whether it is going to cause problems when the CMS or framework is updated six months from now.

AI can make writing the code quicker, but it doesn't take responsibility for the code.

That's still the developer's job.

For me, AI is useful because I already have enough experience to evaluate what it produces, change it where necessary and recognise when its suggested approach isn't a good one.

It's speeding up the process rather than replacing it.

The plugins that solve real problems aren't going anywhere

I don't think AI is going to kill the plugin market.

But I do think it could shrink one particular part of it.

The simple add-ons that do one relatively straightforward thing are suddenly competing against the cost of asking an AI tool to help build exactly what you need.

That is quite a different market to a few years ago.

The plugins that will continue to be valuable are the ones where you're buying more than a few hundred lines of code.

You're buying years of development, testing, documentation, support, compatibility work and specialised knowledge.

Those are still incredibly valuable.

In fact, I think AI might make the distinction between the two even clearer.

For substantial functionality, I'll happily pay someone who has already solved the problem properly.

For the little bits around the edges, though, I'm finding myself reaching for AI before I reach for the plugin directory.

And I suspect I'm not the only developer doing that.