I've had a few prospective search use cases and have used both the percolator out of elastic search and the service out of Google cloud platform. I would find committers that worked on the percolator future and see if your work could possibly be baked into the core elastic search platform. that would solve a lot of people's problems. :-)

and don't forget that solr remains a major player in terms of open source faceted search. and they have no prospective search feature. i would think they would have done it already though - there must be something architecturally that's preventing them from delivering the high-performance solution you built.