neo4j-document-import-skill - Ingest documents into Neo4j knowledge graphs
Ingests PDF, HTML, plain text, Markdown, and semi-structured JSON into Neo4j knowledge graphs with chunking and LLM-based entity extraction.
Tags
Updated: 2026-10-15Capabilities
Typical Inputs
Typical Outputs
What this skill does
- Chunk documents
- Extract entities and relationships
- Load JSON data
- Build document-chunk-entity graphs
- Connect document loaders
- Resolve duplicate entities
- Configure graph schemas
Inputs
- PDF, HTML, plain text, and Markdown files
- JSON or REST API responses
- Neo4j connection details
- LLM or embedding provider credentials
- Graph schema definitions
- Document metadata
- Chunking configuration
Outputs
- Neo4j Document, Chunk, and entity nodes
- Neo4j relationships
- Chunk embeddings
- Entity resolution merges
- Pipeline run results and errors
Requirements
- Python >=3.10
- neo4j-graphrag package
- neo4j driver >=5.17.0
- Neo4j >=5.18.1 or Aura >=5.18.0
- Configured LLM or local Ollama model
- Embedding provider configuration