As you can see from the GitHub log, work on MX, the cross-platform, cross-language cheminformatics toolkit, is currently centered around adding support for complex structure queries with query atoms and query bonds. Although this is not a SMARTS system, the net result would the same - more control over structure queries.
The reason for this focus has to do with a commonly used, but not always implemented substructure searching idiom. Chemists using a substructure search system, like that to be included in ChemVendor, will want to "block" substitution on an atom by adding explicit hydrogens to arrive at the desired valence. So, for example, if Joe draws benzaldehyde with a hydrogen attached to the carbonyl carbon and submits a substructure search, he expects to see benzaldehydes, but not benzamides or benzoate esters.
Given the right pre-processing of the submitted query structure, this kind of simplifying yet powerful substructure searching idom would be supported through MX.