API

An API, short for Application Programming Interface, is a defined way for one piece of software to send a request to another and receive a response. One program sends a message in a specified format; the other reads that message, does something with it, and sends back a result.

For most people outside software development, APIs are invisible infrastructure. When you tap a button in your banking app and your balance updates, an API ran in the background. When a weather widget shows the current temperature, it fetched that data through an API. They are the connective tissue of modern software.

How it works#

An API works through requests and responses. The program that wants something sends a request to the API, specifying what it needs and passing along any relevant information. The service on the other end receives the request, processes it, and sends back a response containing the result.

For AI APIs specifically, the request typically includes text, the prompt or question you want the AI to handle. The response contains the AI-generated output. This exchange happens over the internet, usually in a fraction of a second. The AI model itself lives on the provider's servers; you never need to download or run it yourself.

Developers access AI APIs using programming languages such as Python or JavaScript. They write code that sends requests, handles the responses, and integrates the results into an application. This is what allows a developer to build a smart document summariser, a customer support tool, or a coding assistant without building the underlying AI from scratch.

Examples#

Customer support chatbot. A small business wants to add an AI assistant to their website to answer frequently asked questions outside office hours. Rather than building an AI model, a developer connects the site to an AI API. When a visitor types a question, the site sends it to the API, receives a generated answer, and displays it on screen. The business gets working AI in days rather than years.

Document analysis tool. A legal practice wants software that can read contracts and flag unusual clauses. A developer builds a tool that reads each document, sends sections of text to an AI API with specific instructions, and collects the results. The practice gets a custom tool built around their workflow, combining their own software logic with AI capability supplied via the API.

Why it matters for business#

APIs are what make AI accessible to organisations that are not technology companies. You do not need to train a model, maintain infrastructure, or hire machine learning researchers. You pay for what you use, scale as needed, and benefit from improvements the provider makes over time.

Understanding APIs also helps you evaluate the AI tools you buy. Many products marketed as "AI-powered" are built on top of APIs from a small number of large providers. Knowing this helps you ask sensible questions about data privacy, reliability, and what happens if the provider changes its pricing.

If you want to see how AI APIs work in practice, the tutorial Python AI Setup walks through making your first API call step by step. You can also explore how large language models work, since they are the technology most commonly accessed through AI APIs today. Browse all learning routes on /paths.