Hey Husain here, cofounder of Modulus Good point, we do just that - storing the explicit schema as structured facts. Relevance is based on similarity threshold

Hey Husain here, cofounder of Modulus Good point, we do just that - storing the explicit schema as structured facts. Relevance is based on similarity threshold of the embeddings of the repo purpose and schema fetching is based on structured facts.