hmmm. DK does this, to some degree, but I never trust an algorithm that's looking at a sub-set of specs to match within some unspecified range, and which doesn't know which specs are important to me in this specific application & which are 'don't care', to suggest equivalents blindly - I always check such recommendations, which is just as much work as getting the specs of the part needing to be replaced, and the specs of the proposed replacement, and doing the comparison manually.
Am I just being old fashioned? Or is this the Maker Way? Maybe it's horses for courses; this is my perspective as a designer - sometimes an alternative suggestion that's near enough even though one parameter is 'lesser' than my reference part is still OK.
But if I'm wanting to fix something, my approach assumes that the original designer did their job properly, but given that I'm replacing an unexpectedly broken component, there's probably a solid argument that that's not the case, and so me taking the specs of the failed component as 'gospel' is just as invalid, and perhaps an algorithm that's (hopefully) going to suggest parts whose specs always exceed those of the one searched for is more appropriate.
It's almost like an "alternative suggester" needs at least a couple of modes: "components whose specs all exceed the reference", and "components whose specs are near (above or below) the reference"...