Understanding APIs and How They Work on FlowFuse #flowfuse #api #industrialautomation

Added: 2026-05-05

[00:00:00]APIs are something you hear about all the time in tech. [music] And if you're interested at all in artificial intelligence or machine learning, you're going to spend a lot of your time working with APIs.

[00:00:09]>> [music] >> Hey there, I'm Chris Bernal. I'm a developer relations advocate here at FlowFuse. What is an API?

[00:00:22]>> [music] >> The way I've always explained APIs is to call it sort of a Rosetta Stone for computers. So, consider for a moment a unit like this combination washer-dryer from Samsung. Now, this system does exactly what it says on the tin. It washes and it dries. But, if you think for a moment what it actually takes to be able to do all those things in a single unit, well, it starts to get really complex. You need temperature sensors, systems to detect humidity. You need a way for this system to talk to all of the other connected systems, let alone the Samsung app, your phone, your computer, your TV, and all of the other systems in that ecosystem. That connectivity is exactly where the API comes into play. An API request is usually composed of a few things. First, you have to have an endpoint. The endpoint is where you're actually exposing the service or the request or the process that is going to be invoked by the request itself. Then, you have a method, something like get, post, update, or delete. Then, you have headers and authentication, and sometimes a body that contains a bunch of data or even instructions. If you're making an RPC request over an API, for instance, you can send the request for what you actually want done in the remote system in the body of your request itself. Let's take a moment to specifically define these terms.

[00:01:38]An endpoint is where you send your request.

[00:01:41]The method [music] describes the action.

[00:01:44]Authentication proves that you are who you say you are, and authorization proves that you're allowed to do what you're trying to do. The body can include details of the service needs in order to do the job or specifies [music] what the job itself is. Once the API is done doing what it needs to do, it sends you a response. And that response can [music] be data, status information, or even an error if something goes wrong.

[00:02:04]These responses often come with codes that allow you to actually debug and troubleshoot the responses when you're not expecting to get back what you got back. Anyone who's been on the internet for a long enough time has probably seen the error 404.

[00:02:15]>> [music] >> And that's a status code that tells you that the resource you're looking for doesn't exist on that server.

[00:02:25]Now, here's something that you really need to understand specifically about AI. AI is not ChatGPT. It's not Gemini, it's not Claude, it's not Nano Banana or any of the other systems. It's actually a bunch of systems that are working in a coordinated fashion to process data based upon their training set. And that request, that processing, everything that goes into an LLM that gets you from your request to your response, well, that's all powered by an API. When you send your request from your browser, that's an API. When the model starts processing across distributed courts, it uses APIs. When you get your answer back, that uses an API. When you render that data in a FlowFuse flow on the browser, that uses multiple APIs. In your day-to-day life, you probably use hundreds of thousands of APIs. But, AI specifically requires the use of APIs to get the most value out of your systems.

[00:03:17]What that functionally means is that if you understand [music] APIs, you understand the core building blocks of modern AI products.

[00:03:23]>> [music] >> And whether you're connecting to a chatbot, an image generator, a CRM, a database, or an industrial workflow, APIs are the bridge and the glue that makes it all [music] work.

[00:03:38]FlowFuse is built for users using APIs.

[00:03:41]Our HTTP nodes allow you to interact with any endpoint you can possibly imagine. If the endpoint is accessible and it does something you want it to do, you can connect to it through FlowFuse.

[00:03:50]FlowFuse gives you complete control over the entire data pipeline, and that includes leaving the system as well as bringing data back [music] in. That gives you total freedom, total flexibility, and extensibility. We hope you like this video. If you did, please consider giving it a like, comment, or subscribe. If you'd like to learn more about AI topics just like this one, then check the link in the description below to register your interest for the upcoming FlowFuse Academy. And if this piqued your interest in FlowFuse, head on over to flowfuse.com to start your free 14-day trial. This has been Chris Bernal. Thanks for watching, and I'll see you next time.