A crawler visits a website just like a visitor. It reads the text line by line, follows every link and recognizes what matters: product names, prices, addresses or dates.
It plays by the rules of the website: whatever the robots.txt blocks, it leaves out, and it sends requests so gently that the server never breaks a sweat.
With modern technology, such as a real browser in the background and natural pauses, it behaves like a person at a screen and is hard to tell apart from a normal visitor.
Whatever it finds, it stores in an orderly way, for example as a spreadsheet or straight into your database. The next day it comes back and reports only what has changed.