Ask HN: How can I find out what a website is actually about?

For a project of mine we want to analyse what a website is actually about. We need to obtain a list of keywords what the website is about.

Right now we are just using meta tags to get a list of keywords. But this does not offer very good results.

How does Google do it? They must have some way to find out what a website is actually about.

Do you know external APIs that could help?

Thanks.