I've got several bots that scrape various places for transient information which isn't stored anywhere but merely transformed and then posted to the Fediverse and/or notified to my phone.

(I suppose you could argue that the information ending up in a post or notification is "non-volatile" storage but honestly? People will laugh at you.)

Price comparison. Alternative front ends.

It's irrelevant to the definition of "scraping." Scraping from a website is grabbing data, in bulk, not through the website's own interface. It doesn't matter if you save the data, use it in RAM, or just download directly into `/dev/null`. Just like scraping paint off of a wall is still "scraping" whether you're letting it fall onto the floor, collecting it into a trashcan, or sweeping it out the door afterwards.