Purely theoretically: on what principle do online aggregators like mamondo, skyscanner, jooble, etc. work?

Question

Parsing

Friends, a purely theoretical question, how do online aggregators work? By the principle of parsing data from other sites?
Or do they cooperate with end sites and connect to their databases to display information?
Or is some other method used, for example, they send a request to search engines and parse the result of the search results in search engines, and not from specific sites? Is there anyone who knows?
Always interested in this question. After all, parsing a specific site is difficult, they can change the layout and that's it, change the algorithm again, if I'm not mistaken of course. Maybe someone took part in the development of such aggregators, please share information on what principle most online aggregators work on?
Interested in the technical side of the issue, the principle, without details.

Answer 1

They have contracts, of course, because they need to receive a commission from sales from this

Answer 2

In general, the essence of the work of most aggregator sites, as I understand it, is the following, the contract, the aggregator is given access to data in the form of XML, and so on. Principle, we have traffic, you have a product - a service. They operate with data, and I give links to the source site. But this is rude.
I thought they just parse the data, scan the pages of sites and independently give out information. No, everything is in agreement.