It's exactly that. Redis is an in-memory data structure server that you can outsource index-style operations to. Vector similarity is a type of index search. I think it's an exact fit for Redis.