Is there an easy way to update a schema without reloading all data (eg: changing a field to a facet)?

Not at the moment unfortunately, but updating a collection's schema is on our roadmap. In the meantime, I'd recommend using aliases (similar to symlinks in linux) to point to a timestamped collection and using that alias for searches. When you need to update the schema, you'd create a new collection with the latest timestamp in its name, index the data, and then finally switch the alias to point to the new collection.

ah, cool! Only question is what happens once I have 100k indexes? Can I purge them?

Once you update the collection alias to the latest one, you'd finally want to drop the old collection

Gotcha, this is fantastic, thanks