Turn documents into high-quality instruction datasets with grounding, quality filtering, deduplication, and provenance. - View it on GitHub