I think the only flaw here is that while it may be trivial to get raw data from say fossil, any major multi-department retailer (Amazon, Walmart, Best Buy, Target, B&H...) is not going to be this easy.

Prices are competitive information, at least competitive enough to warrant obfuscation and preventing scraping.

And, the more sources you have, the more likely things will constantly fail. Scraping is always fragile.

One of the first web-scraping projects I ever built was to find the lowest price for Clif bars. I ate about 365 per year, and in only 45 minutes of programming a cron job and push notifications, I easily saved $100 that year.

As someone who may or may not monitor very elastic competitor pricing (allegedly) I would add that if you are not respectful with your scraping you will get IP banned. You might also get shut down by a provider (e.g. aws).

Services like Crawlera act as a proxy that also forces you to be nice. If anyone is getting into this, I would recommend they use a service like that to scrape at scale.

Honestly I'd prefer if the process didn't even have to "know" sites.

Let me just supply the URL and the price string... and let it automatically figure out the right <div> or class or id or surrounding text it can be extracted from. And if it ever breaks (can't be found, or returns something in a different numeric format) then just notify me to visit it and see what's up.

Command-line is one way to do it (can be the canonical version), but then a browser extension makes it even easier. You'll need a headless browser even from the command-line in any case, since the product info isn't always in the HTML at all.

Sounds similar to what Klaxon does. https://github.com/themarshallproject/klaxon