Hey there, by semantic we mean, embedding text and storing it in the vector DB. Regarding scale, we thought a lot about it, and that's why we process span in th

Hey there, by semantic we mean, embedding text and storing it in the vector DB. Regarding scale, we thought a lot about it, and that's why we process span in the background queue. Tradeoff would be that indexing / embedding would not be real-time as scale will grow. We will also use tiny embedding models, which become better and better.