neo4j-document-import-skill

What problem does it solve?

Neo4j document ingestion typically requires significant manual effort to chunk content and model extracted entities and relationships; this Skill guides agents through converting PDFs and text into a connected knowledge graph for graph and GraphRAG-style workloads.

Core Features & Use Cases

• Document chunking to graph nodes: Splits input documents into :Document and :Chunk structures suitable for downstream retrieval and traversal.
• LLM-driven entity and relationship extraction: Uses SimpleKGPipeline (neo4j-graphrag) and related components to turn text into typed entities/edges.
• Flexible ingestion options: Supports semi-structured JSON loading via apoc.load.json, and provides a no-code alternative via Neo4j LLM Graph Builder.
• Graph schema & resolution guidance: Helps define a schema, run entity resolution, and set constraints/indexes for reliable ingestion.

Quick Start

Tell the agent to ingest a file like report.pdf into Neo4j using SimpleKGPipeline with a defined GraphSchema, chunking, embeddings storage, and post-ingestion entity resolution.