Currently any field on the index will be appended to the rendered content as well. This means that fields included as metadata will pollute the rendered content and thus the Vragen.ai knowledgebase. We should remove this functionality and only include the rendered content field. Another way to fix this would be to differentiate between fields that can get embedded (are part of the content) and fields that are metadata. We need to be careful about backwards compatibility with any changes we make.