A worthy use case to share here. Not one of ours, but a strong example of building Neo4j-style knowledge graphs from unstructured text. Lettria used hybrid embeddings + Qdrant vector search + a RAG pipeline to extract subject-predicate-object triples and structure them into a graph.
GraphRAG: How Lettria Unlocked 20% Accuracy Gains with Qdrant and Neo4j - Qdrant