How viable would it be to implement RegEx searching into Kagi? As thekingofravens put it, "[...] depending on how the engine was made, this could range from easy to very difficult to implement", which is an important factor to consider. If it turns out it would be very difficult and time consuming then it might not be worth investing the time into, but if not then I think it's definitely something worth at least looking into.
It would be an incredibly powerful tool to have when looking for exact matches to a range of possible values known to follow specific formats. For instance, if you know the documents you're looking for will contain an ID string in the format of [A-Z]{3}[0-9]* (ie, ABC1234, BBQ7, etc) and you're only interested in A[BC]{2}[0-9]{4} (ie, ABC1234 but not BBQ7), then being able to use RegEx would make short work of such a search, accurately and predictably filtering out the unwanted results and only leaving exact matches.
I'm not aware of any search engine that currently provides this kind of fine-grained and predictable control over searches. Besides strong pattern matching it would allow for exact searches of characters, rather than ignoring them. I'm not sure how Kagi treats special characters, but I know Google for instance ignores a lot of special characters in queries, resulting in unpredictable results for me.
I understand that exact matches are not always desirable (and regex would definitely confuse those not familiar with it), hence why this would need to be a per-search opt-in feature, but I also think it would be nice to know that the query is being processed exactly as user-intended, rather than it being a black box machine that may or may not ignore parts of the query, resulting in undesired items ending up in the search results.
RegEx-powered search would be a very powerful tool to add to Kagis arsenal.
