I’ve been working on a search concept for Airtable that has kept me awake a few nights.
One of the [many] problems with universal search (i.e., a search that spans multiple bases and multiple tables) is the complex types of fields and content you’ll find when the entire galaxy of data possibilities will be encountered. You have to embrace all that Airtable could ever store and in a way that is not only findable but indexable in a reliable and compressed manner.
Another problem; Airtable is fun to use. Users don’t want to search in a completely disengaged web page away from their data. If you’re at all like me, you probably want your search results displayed like any other information objects are displayed in the platform you’re searching.
I struggled with many aspects of an open (but secure) and easy-to-understand architecture. One of the biggest aspects of this was answering a basic question - if you had an index of all your data, where would you feel comfortable storing it? The answer is simple - why Airtable, of course. So I did that. Airborne indexes your content into your Airtable workspace where it is stored and available to use in a variety of use cases including an embed block (where I demo it in a video).
Lastly, relevance is pretty important; you don’t just want a list of hits. You really want a list of hits ordered by some sort of ranking score. For example, consider a keyword search for “sales projections”. If this appears in a base name it’s probably more relevant to your quest than when it appears in a field.
These basic requirements seem simple at first glance.
In any case, I threw some mud at the wall and created Airborne Search - a very early prototype of what I hope will provide me, my clients, and possibly all of you with search alternatives in addition to table-level filters and such.
Take a look. Tell me what you think. I’m looking for some people to test it a little.