neo4j-vector-index-skill

What problem does it solve?

This Skill helps you turn text (or other signals) into embeddings and then create and use Neo4j vector indexes to perform semantic similarity search over your graph data.

Core Features & Use Cases

• Vector index creation & tuning: Create and configure vector indexes (dimensions, similarity function, HNSW, quantization) and wait until they are ONLINE before querying.
• Embedding ingestion pipelines: Ingest embeddings either via batch ingestion (UNWIND + setting vector properties) or via in-Cypher embedding using ai.text.embed(), including dimension-safety checks.
• Vector retrieval & hybrid retrieval: Query nearest neighbors using the SEARCH clause (Neo4j 2026.01+) or the db.index.vector.queryNodes() procedure fallback, then optionally combine results with graph traversal or other ranked sources.

Quick Start

Create a Neo4j vector index for chunk embeddings, ingest embeddings with the same model and dimensions, then run a vector similarity search using the SEARCH clause for your query embedding.