I would distinguish between parsing and scraping. Parsing really needs a, well, parser. Otherwise you’ll get things wrong on perfectly well formed input and your program will be brittle and weird.

A scraper is already resigned to being brittle and weird. You’re relying not only on the syntax of the data, but an implicit structure beyond that. This structure is unspecified and may change without notice, so whatever robustness you can achieve will come from being loose with what you accept and trying to guess what changes might be made on the other end. Regex is a decent tool for that.