You can write a piece of software to:
- Take the keywords from the regular expression;
- Google the keywords and get a list of results;
- Crawl each resulting URI and filter it with complete regular expression.
Let's study a case: from
site:gog.com find all games that have Spanish voice-over.
The regular expression is:
It shall match, for example:
Audio lanuage: English, German, Spanish, French.
And not match:
Audio language: only-English. Text language: Spanish.
Step 1. Let your software search this on Google:
site:gog.com audio Spanish inurl:game
inurl:game here means only search in game description pages
Step 2. Get the 300 resulting links and crawl into every one of them.
Step 3. Filter the result with the given regular expression:
This should be easy to build. In fact I don't understand why I couldn't find something that is already built that way.
Since search engines can't afford the resource to scan their data with regular expression, this dirty job falls on your part, and your computer should do that with what search engines already provide.