While that may be sort of what you are looking for, it is somewhat lacking in that the results pages are generated with quite lousy HTML. It might do for the short term though.

You can generate fully customizable results with the Google API, but for that you’ll need to do a little scripting, and get a Google key for API access ( http://www.google.com/apis/ ).

There are a couple of examples of embedded search and some example scripts at the following:
At fiftyfoureleven.com
At Dankarran.com (I really like the search results of this one, he uses it on his search page for the site.)

I’m sure there are others floating around, but those two have some full examples given on how to do it.

Firstly, I know that Google’s results are not going to be as up-to-date as a search on WordPress database.

Secondly, while the results work for me, the display of the search results is not perfect – in other words, it looks like a Google search page, it’s not integrated with the rest of the site design.

Having said that, it does effectively return results whether they are WordPress-powered content (for the news updates) or content stored on the server in static pages (basically everything else).

Is there any way I can customise the Google search results page any further? (I’m sure I’ve seen a Google search results page feature a site’s icon – but perhaps that’s a premium, paid-for service).

On the other hand, I see absolutely nothing wrong with using a Google search entirely, so long as you are getting good results.

Personally, I rarely use site searches, because:
A) I find the results are often weak or dubious,
B) site based search engines I feel are relatively slow,
C) I’m a horrible speller and site searches rarely have decent spellchecking, and
D) I can more easily expand my searches to the entire web just by removing site: for instance.

Of course these aren’t true everywhere, but I’ve had enough bad experiences on even well funded and staffed sites, that I rarely even look at site search boxes (one exception being Wikipedia, which I use despite having most of the flaws I’ve listed here–I guess it has to do with placement–often I resort to Google eventually).

I never understood the desire that publishers (of which I must admit, I am not) had with running and maintaining their own search databases and engines. It seems like a lot of overhead for something that’s really, really hard to do well.

~~Ian

The best of all is, it works using a spider in a webinterface, and you can automatically add content to it on-the-fly. There’s also extensive support for not indexing some parts of pages, templates, query stats, spidering logs, etc.

Runs smoothly on 2 of my sites.

http://www.cs.ioc.ee/~ando/sphider/

Personally, I hate half-measures such as some posts in a database, and some not.

There are some solutions to your enquiry but they are as elegant as my suggestion that all posts should be database-driven.

All you gotta do is just make time to migrate all posts to the WP database. It’s all worth it for the search, but up to you

There is one downside though – the way results are displayed in google’s public service search uses the same code as google’s normal results and is really rather rubbish… this, that…