Build a robots.txt without memorizing the syntax

At the root of your domain - https://example.com/robots.txt. It won't be read anywhere else, including subdirectories.

No. Disallow tells well-behaved crawlers not to fetch a page, but if other sites link to it, Google can still index the URL without visiting it. Use a noindex meta tag on the page itself if you need a guarantee.

That's a real, current judgment call, not something with one right answer - blocking them keeps your content out of that company's training data, but also out of AI-powered search/answer surfaces that increasingly send referral traffic. Decide per bot, don't block by default.

An empty Disallow: with no path means "disallow nothing" - it's equivalent to allowing everything for that user-agent, and is valid, intentional syntax for an "allow all" block.