@curiospumpkin Hey! 👋

We actually do support custom embedding models, but that’s part of our Enterprise Plan. Changing the embedding model means retraining the whole chatbot, so it’s not something we expose by default to make it simple for users and avoid any issue.

Right now, we’re already using Voyage 3.5, one of the best embedding models out there. But for enterprise customers, we offer customising embedding options too.

Noted on this Rohit, I’m assuming the same for rerankers as well?

Yes, we do support different ranker as well.