Building Agents in Neo4j Aura
Build agents on Neo4j Aura without code. Design an agent, add Cypher Template, Text2Cypher, and Similarity Search tools, then publish it over MCP.
In this 2-hour course, you will learn
Building Agents in Neo4j Go to glossary for Aura (opens in a new tab)Neo4j's fully managed cloud service. is an intermediate course that teaches you how to create AI agents directly in the Aura Console, without writing application code. In around two hours you will design an agent, equip it with 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. Template, Text2Cypher, and Similarity Search tools, and publish it over MCP so external AI applications such as Cursor can call it.
Aura Agent turns natural language questions into graph queries that run against your data in Go to glossary for AuraDB (opens in a new tab)The Aura product for transactional workloads.. Instead of wiring together an 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., a driver, and a retrieval pipeline yourself, you configure the agent visually: you describe its role, give it tools, and test it in the Console. Because every answer is retrieved 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. rather than generated from the model's training data, the agent stays accurate about your domain.
Aura Agent
Understand what an Aura Agent is and how it orchestrates tools to answer natural-language questions against a Neo4j graph database without writing application code.
Agent Design and Tool Selection
Design an agent by defining its purpose, selecting appropriate tools, and configuring the system prompt to control how the agent interprets and routes user requests.
Cypher Template and Text2Cypher Tools
Build a Cypher Template tool for predictable parameterised queries and a Text2Cypher tool that translates natural-language input into executable Cypher at runtime.
Similarity Search
Configure a Similarity Search tool that uses vector embeddings to retrieve semantically related nodes from your graph based on a user's natural-language query.
Publishing Agents with MCP
Publish a completed agent over the Model Context Protocol so external AI clients and LLM-powered tools can discover and call it as a reusable MCP resource.
3 modules, 2 hours.
Who this course is for
This course is for anyone who wants to build a graph-backed AI agent without writing and hosting application code: developers prototyping retrieval over 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. before committing to a custom stack, data practitioners and architects who know their graph well but do not build applications day to day, and product-minded builders who work in the Go to glossary for Aura (opens in a new tab)Neo4j's fully managed cloud service. Console. You need an Go to glossary for AuraDB (opens in a new tab)The Aura product for transactional workloads. instance with data loaded and Aura Agent enabled for your organization, and you should be comfortable in the Console and able to read and adapt a 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. query - Aura Fundamentals and Cypher Fundamentals cover this ground. You do not need experience with LangChain, agent frameworks, or MCP; the course introduces the agent concepts it uses as it goes.
What you'll do
This is a hands-on course built around challenges. You will create a real agent in your own Go to glossary for Aura (opens in a new tab)Neo4j's fully managed cloud service. Console and improve it step by step. You will start by generating an agent with AI from a prompt and your instance schema, then extend it yourself: writing a 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. Template tool with named, typed parameters, adding a Text2Cypher tool for ad-hoc questions where the query cannot be known in advance, and learning when a Similarity Search tool is the right choice for vector-based, semantic lookups. Along the way you will test the agent with real prompts and refine the role, scope, and tool descriptions that steer its reasoning.
Finally, you will publish the agent: comparing Internal and External access modes, enabling the MCP server, and connecting the agent to Cursor using its MCP endpoint, so an external AI application can call your graph as a tool.
Where to go next
If you want to understand the concepts behind the agent you just built, Neo4j and GenAI Fundamentals covers grounding, retrieval, and how graphs improve 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. accuracy. To go beyond the no-code tools, take Developing with Neo4j MCP Tools, then build your own MCP tools in Python or TypeScript. If your graph does not exist yet, Building Knowledge Graphs with LLMs shows you how to build one from unstructured data.