An API is a defined set of rules that lets one piece of software communicate with another, allowing applications to request and exchange data in a structured format, such as JSON, without needing to parse raw HTML. When a site loads its data from an API, Scrapy's documentation calls reproducing those requests the preferred approach, since it gives structured, complete data with minimum parsing time.
Some proxy and data-extraction providers offer their own "scraper APIs" that combine proxy infrastructure with built-in parsing, effectively turning an unstructured website into a structured API response on the customer's behalf.
A public API can document its rate limits and authentication requirements. GitHub's REST API documentation, for example, publishes its hourly request limits.