Thanks for the quick response! We are currently performing postfiltering. We query Postgres with user-supplied filters, and then combine that with the response from the LookupResources call to further filter the results based on the returned IDs. Given that we're postfiltering, does your recommendation to store ZedTokens alongside the data still apply, and how would that work in practice? We were envisioning a table-level ZedToken. For example take the "files" table. The idea we had was to pass this token to the LookupResources call within our RLS policy. The ZedToken would essentially represent the last time which the files table was updated. Does this work, or do I have a fundamental misunderstand of how ZedTokens work? Would this provide sufficient consistency, or is there a specific reason why we'd need a ZedToken for every file? If we were to store a zed token for every file in the files table how would we use those tokens to query the lookupResources api?