This Week in Neo4j: NODES, StrangerGraphs, Claude Skills, MCP and more

Photo of Alexander Erdl

Alexander Erdl

Senior Developer Marketing Manager

3 min read

Aleksandr Khazov

Welcome to This Week in Neo4j, your fix for news from the world of graph databases!

We dive into HopperGraph, a “Stranger Things” graph, where you can explore character relationships and surface data-driven Season 5 predictions straight from the Upside Down.

This edition also brings every NODES 2025 session to YouTube, shows how to extend Claude with Neo4j-aware Skills, and demonstrates how to build powerful graph agents in n8n using Neo4j MCP.

And while we have just finished NODES 2025, we are already preparing for the next one. The Call for Papers for NODES AI is open: On April 15, practitioners will gather virtually to explore AI, context engineering, and intelligent agents.

Happy Graphing,

Alexander Erdl

 

COMING UP!

As the founder of LEXRAG, Aleksandr Khazov utilises the GraphRAG architecture with Neo4j at its core to enable precise retrieval and reasoning over complex legal texts. His work integrates structured legal ontologies, vector embeddings, and LLMs to power legal research assistants.

Connect with him on LinkedIn.

Aleksandr delivered one of the top-rated sessions at NODES2025: “GraphRAG for Law: Building Legal Reasoning Agents with Neo4j and LLMs”. He walked us through the process of modelling granular legal entities as graph nodes, enriching them with embeddings, and orchestrating hybrid retrieval that combines semantic search with deep legal structure.


Samira Korani

 

NODES: Watch Sessions


Missed NODES 2025 on November 6? All the graph action is now on YouTube. Every session from our AI Engineering, App Dev, Data Intelligence and Knowledge Graphs tracks is being uploaded to the official NODES 2025 playlist, so you’ll be able to watch them all on demand.

 

STRANGERGRAPHS: “Stranger Things” - The Gate to True Sight With Graph Intelligence


Stephen Chin invites us to discover HopperGraph, a Neo4j-powered “Stranger Things” knowledge graph built from 150k Reddit fan theories (234k nodes, 1.5M relationships) that lets you explore character connections and surface graph-driven predictions for Season 5 – including the fate of Hawkins, Eleven and Will. Use the Stranger Graphs experience to navigate the Upside Down like a graph and test your own finale theories against what the data suggests.

 

CLAUDE: Using Claude Skills with Neo4j


This article from Tomaz Bratanic explores how to extend Claude with “Skills” modules to enhance its interaction with Neo4j: teaching Claude to generate up-to-date Cypher queries and invoke Neo4j via the Model Context Protocol (MCP). The method enhances LLM-driven graph workflows by making them more accurate, scalable, and context-aware when working with graph databases.

 

MCP: UNLOCK the Power of Graph Agents with Neo4j and n8n


In this video by Daniel Walsh, we are shown how to build a Neo4j-backed knowledge graph and connect it to an n8n AI agent using Neo4j MCP and Claude Desktop – no Cypher is required. The video walks through two concrete graph agent use cases: a Customer 360 graph and a document-structure graph for smarter, more agentic retrieval of complex legal documents.

 

 

POST OF THE WEEK: Thomas L.



Please share it if you like it!

Share Article

Explore

Samira Korani

This Week in Neo4j: NODES, Retrievers, Graph Analytics, Agents and more


4 min read

Why Startups in the Google Ecosystem Choose Neo4j


4 min read

Andrew Ng

This Week in Neo4j: Context Engineering, Chunking, Cypher, Aura and more


3 min read

A constellation of Orion represented as a graph with nodes and relationships.

Advanced RAG Techniques for High-Performance LLM Applications


17 min read

Analyze Graphs Directly in Snowflake — The Wait Is Over


14 min read

Isaac Ritharson

This Week in Neo4j: Aura Agents, Fraud, Langchain, AI Memory and more


4 min read