The random UI change(s) is something we have to detect really quickly & fix - or maybe in the future use AI models to read the data (if there was a way to ascertain that there's be no hallucinations).

Per the ToS and scraping, the way the SDK is built is akin to if we had a browser like Brave, Gener8 or even Chrome itself. These browsers essentially already "scrape" what the user is doing.

We think end-users should & will always have the right to always decide how they'll browse the web and can share whatever they choose to.

It's also similar to installing a chrome extension that does the "scraping" for you - just with a much smoother UX, we think the end-user should always have a right to decide to do stuff like this.

Lastly, we stay clear of any intellectual property, the only data we collect on behalf of the user is data on activity that they generate - so for example, Gandalf can tell the title of the show a user watched on Netflix, but we do not scrape images, video content, anything that's Netflix's IP.