GraphRAG Hackathon
Hands-on GraphRAG hackathon: model a knowledge graph in Neo4j Aura, then build your own GraphRAG application with an AI coding agent and MCP servers.
In this 43-minute workshop, you will learn
The Full definition for GraphRAG (opens in a new tab)Retrieval-augmented generation whose context comes from a knowledge graph, so the model can follow the relationships between facts. Hackathon is a hands-on Neo4j workshop with a difference: after a short introduction to GraphRAG, graph modeling, and Full definition for Cypher (opens in a new tab)Neo4j's implementation of GQL, the ISO standard query language for graph databases. It is declarative: you describe the pattern to find, and the database decides how to find it., you spend the rest of the session building your own GraphRAG application with the help of an AI coding agent connected to the GraphAcademy MCP server.
The course is designed around a live hackathon session - taking part earns you a certificate of attendance - but the materials stay available afterwards, so you can also work through it at your own pace.
GraphRAG
Understand how Graph Retrieval-Augmented Generation combines knowledge graphs with large language models to produce more accurate, context-aware AI responses.
Neo4j
Build and query a knowledge graph in Neo4j Aura, modelling real-world entities and relationships to power a production-ready GraphRAG application.
Cypher
Write Cypher queries to traverse graph relationships, filter nodes by property, and retrieve structured context that feeds directly into your AI pipeline.
MCP Server
Configure and connect an MCP server so your AI coding agent can query the Neo4j graph at runtime, grounding its responses in live, structured data.
AI Agents
Apply an AI coding agent to accelerate GraphRAG development, directing it to generate, debug, and refine code against a running Neo4j instance.
What you need to take part.
A GraphAcademy account
The workshop is delivered here on GraphAcademy, so you need to be signed in to work through the lessons and keep your progress. Creating an account is free.
A GitHub account, or Git on your own machine
You write code against the workshop-hackathon repository. The quickest route is a GitHub Codespace, an online editor that clones the code and installs everything for you — that needs a GitHub account.
If you would rather work locally, clone the repository with Git and run it in your own editor instead. No GitHub account is needed for that.
3 modules, 43 minutes.
Who this workshop is for
This hackathon is for developers. If you have built applications with Go to glossary for large language model (opens in a new tab)A model trained on text to predict the next token, and so to generate language. APIs and want to see what a Full definition for knowledge graph (opens in a new tab)A representation of real-world entities and their relationships, stored according to organizing principles, typically in a graph database. adds - grounded answers, explicit Full definition for relationship (opens in a new tab)A named, directed connection between two nodes. Every relationship has a type, a start node and an end node., and queries an LLM can reason over - this is a fast, practical way to find out. It also suits anyone curious about agentic coding, since your AI coding agent does much of the implementation work. No prior knowledge of Neo4j or graph databases is required - you need only a GitHub account for GitHub Codespaces and basic programming experience in Python or JavaScript/TypeScript. The workshop provides temporary access to a hosted language model, so you do not need your own LLM provider API key.
What you'll do
You start by getting your environment ready: create a Neo4j Go to glossary for Aura instance (opens in a new tab)A single Neo4j database running in Aura., tour the Go to glossary for Aura (opens in a new tab)Neo4j's fully managed cloud service. console, and connect your coding agent to the Aura and GraphAcademy MCP servers so it can create data models, generate mock data, and run Full definition for Cypher (opens in a new tab)Neo4j's implementation of GQL, the ISO standard query language for graph databases. It is declarative: you describe the pattern to find, and the database decides how to find it. on your behalf. The introduction then covers the essentials in about thirty minutes - why graphs make Go to glossary for large language model (opens in a new tab)A model trained on text to predict the next token, and so to generate language. smarter, how Neo4j stores Full definition for node (opens in a new tab)A vertex in a graph. In a property graph it can carry labels and properties. and Full definition for relationship (opens in a new tab)A named, directed connection between two nodes. Every relationship has a type, a start node and an end node., how to model your domain as a graph, and how to read data with Cypher - before you design a Full definition for data model (opens in a new tab)The labels, relationship types and properties chosen to represent a domain. for your own idea and generate mock data to populate it.
Then it is your turn to build. A walkthrough shows you the shape of a working Full definition for GraphRAG (opens in a new tab)Retrieval-augmented generation whose context comes from a knowledge graph, so the model can follow the relationships between facts. application, and from there you and your coding agent build your own: it takes a natural language question, retrieves relevant context from your Full definition for knowledge graph (opens in a new tab)A representation of real-world entities and their relationships, stored according to organizing principles, typically in a graph database., and uses an LLM to generate a grounded answer. Because the heavy lifting is delegated to your coding agent, the session is less about typing code and more about deciding what your graph should look like, which questions your application should answer, and whether its answers are genuinely grounded in your data.
Where to go next
The hackathon moves fast, so the natural next step is to fill in the foundations: Neo4j & Generative AI Fundamentals covers Full definition for GraphRAG (opens in a new tab)Retrieval-augmented generation whose context comes from a knowledge graph, so the model can follow the relationships between facts. concepts in depth, and Neo4j Fundamentals plus Cypher Fundamentals consolidate the graph and query skills you used during the build. If the MCP integration caught your interest, Using MCP with Neo4j Tools shows you how to connect AI agents to Neo4j in your own projects, and Graph Data Modeling Fundamentals helps you revisit the Full definition for data model (opens in a new tab)The labels, relationship types and properties chosen to represent a domain. you sketched during the build. For more instructor-led sessions like this one, see the workshops page.