There's a word I keep hearing whenever I go to an AI hackathon.

“API.”

Of course, I've heard it plenty of times before. I even build apps with AI myself.

But if someone asked me, “So, what exactly is an API?” could I explain it properly?

Hmm. Maybe not! Haha.

I'd always thought of it as something like a connector that links one thing to another.

That was about the extent of my understanding.

At the Tokyo Agent Hackathon on Saturday, October 17, I heard “API” over and over again.

The Next Day, We Started Talking About Amazon APIs!

The following day, I was meeting with a teammate from a different hackathon, Vietnam Japan AI Hackathon 2026, to discuss the AI app we were developing.

And somehow, the conversation turned to connecting an external app with Amazon Seller Central through an API.

Now that was a familiar subject for me!

I sell products on Amazon, so I deal with product information, inventory, prices, orders, and other data all the time.

But whether an external app can directly access that information is a different question.

Amazon has an interface for sellers called the Selling Partner API, or SP-API.

With the appropriate APIs and permissions, apps can retrieve information such as orders and inventory, or manage certain listing information.

Of course, that doesn't mean every possible action is available.

What you can and can't do depends on the specific API, your permissions, and Amazon's rules.

So we started discussing which actions could actually be handled through an API.

And then the conversation took a funny turn.

Repacking? An API Can't Do That by Itself! Haha!

One of the things that came up was “repacking” or “repackaging.”

In other words, physically taking a product out of one package and putting it into another.

Well, an API isn't going to reach into a box, take out the product, and pack it again!

We both had a little laugh about that.

Of course, it's possible to imagine a system that connects to a warehouse and sends instructions for repacking work.

But exchanging data and physically handling a product are two different things.

Having an API doesn't magically automate everything.

It sounds so obvious, but thinking about a real-life example like this made the role of an API much easier for me to understand.

It also reminded me that you have to check what each service's API actually supports.

So, What Exactly Is an API?

API stands for “Application Programming Interface.”

Hmm. Even knowing the full name doesn't really help a beginner like me! Haha.

Put simply, an API is a set of rules and an interface that lets software exchange information or use certain functions.

For example, suppose I want to show Amazon inventory data inside an app I've built myself.

My app could send a request through an appropriate Amazon API: “Can you give me the inventory information?”

If the app has the necessary access and permissions, Amazon's system can return the data.

Then my app can display it on the screen.

The flow would look something like this:

My app → Request through an API → Amazon's system → Data comes back → My app displays it

Aha!

Now that makes a little more sense.

So my original idea of “something like a connector” wasn't completely off.

But an API and a connector aren't exactly the same thing.

An API provides a defined way for software to interact with another system.

A connector is often a tool or integration that uses an API—or another method—to connect different services.

A small difference, but an important one!

And What About an API Key?

There's another term that often comes up alongside API.

“API key.”

I hear that one all the time when using AI tools, too.

If an API is the interface used to access a service, an API key is a bit like a key or an ID that helps the service identify the app making the request.

Depending on the service, it can help determine who is accessing the API and whether they're allowed to use it.

For example, when you want to connect your own app to an AI model provider, you might create an account, generate an API key, and add it to your app's configuration.

Then your app can use that key when making requests to the AI model.

But not every API uses API keys.

Some services use other authentication methods, such as sign-in flows or access tokens.

And here's one really important thing:

Don't share your API keys with other people!

With some services, someone who gets hold of your key could use up your account's allowance or credits.

That's why it's important not to put secret API keys in public GitHub repositories, blog posts, or social media posts.

Since I build apps with AI quite often, this is definitely something I want to understand better.

Maybe I've Been Using APIs in My Own Apps Without Even Realizing It?

That got me thinking about something else.

I've built various apps using tools like Claude Code.

But I don't fully understand every line of the code those tools have written.

So maybe I've already been using APIs without really noticing!

For example, an app might retrieve information from an external service.

Or send text to an AI model and receive a response.

Or show data from another system inside its own interface.

APIs can be part of how these things work.

Of course, APIs aren't the only way to connect systems, and I can't know which APIs my own apps actually use without looking at their code and settings.

Next time Claude Code helps me build an app, maybe I'll ask:

“Does this feature use an API?”

I have a feeling I'll start noticing things I never paid attention to before.

APIs Are Starting to Feel a Little More Familiar!

On Saturday, I kept hearing the word “API” at an AI hackathon.

On Sunday, an Amazon-related conversation got me thinking about what APIs can and can't do.

And now I'm beginning to understand what an API key is, too.

I still can't write an entire API integration from scratch by myself.

But one of the fun things about building apps with AI is how a technical term you barely understood can suddenly connect to your own work or everyday life.

That's exactly what happened to me with APIs.

Maybe learning something through real experience is easier than just memorizing a definition.

Next, I'd love to look at the apps I've already built and find out where they're using APIs!

It feels like another little corner of the AI app development world has opened up to me.