GraphAcademy
Course

Build a Knowledge Graph MCP Server Workshop

Hands-on workshop: build a knowledge graph for your use case, query it with Cypher, and expose it to AI agents through your own MCP server.

2 hours17 lessons across 4 modules
About this workshop

In this 2-hour workshop, you will learn

The Build a Knowledge GraphA representation of real-world entities and their relationships, stored according to organizing principles, typically in a graph database. MCP Server Workshop is the companion course for the instructor-led workshop on giving AI agents access to your data through a knowledge graph and an MCP server. Over roughly two hours, you build a knowledge graph for your own use case, write the queries that answer your questions, turn those queries into MCP tools, and connect an agent to them.

Agents are only as good as the tools they use to make decisions, gain context, and take action. A knowledge graph gives an agent structured, connected data, and MCP gives any agent access to it.

  • Knowledge Graphs

    Understand what a knowledge graph is and why agents benefit from structured, connected context, then model your own use case and add sample data to it.

  • Querying with Cypher

    Read a graph with Cypher patterns and write read-only queries that answer the questions your use case needs answered.

  • MCP Server Tools

    Set up a scaffolded MCP server and turn your queries into parameterised tools that give an agent access to your knowledge graph.

  • Agents and MCP

    Connect an agent to your MCP server, then test how it chooses and uses your tools to answer questions about your data.

Before you start

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.

    Sign in or create an account

  • A GitHub account, or Git on your own machine

    You write code against the workshop-mcp 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.

Course content

4 modules, 2 hours.

  1. 5 lessons
    1. Workshop Overview5 min
    2. What is a Knowledge Graph5 min
    3. Model Your Use Case5 min
    4. Add Unstructured DataOptional5 min
    5. Add Sample Data5 min
  2. 4 lessons
    1. Cypher Patterns5 min
    2. Matching, Filtering, and Returning5 min
    3. Aggregating, Ordering, and Limiting5 min
    4. Query Your Use Case5 min
  3. 6 lessons
    1. Set Up Your Development Environment5 min
    2. What is MCP5 min
    3. Review and Test the Scaffold5 min
    4. Define Your Tools5 min
    5. Create Your Tools5 min
    6. Add a Text-to-Cypher ToolOptional5 min
  4. 5 lessons
    1. What is an Agent5 min
    2. Connect Your Agent5 min
    3. Experiment with Your Tools5 min
    4. Connect a Python AgentOptional5 min
    5. Next Steps5 min

  • Who this workshop is for

    This workshop is for developers and technical practitioners who are new to Neo4j and want to give agents access to their own data. You should know what a graph database is and be able to read Python programs. You will also have a coding agent to help with implementation details. It is designed to be followed live with an instructor, but every lesson includes what you need to work at your own pace.

  • What you'll do

    You start by describing your own use case and the questions a knowledge graphA representation of real-world entities and their relationships, stored according to organizing principles, typically in a graph database. needs to answer for it, then model it and add sample data. You learn the CypherNeo4j'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. needed to read a graph and write a query for each of your questions.

    You then set up a development environment, build on a scaffolded MCP server to create a tool for each question, and test the tools. Finally, you connect an agent to your MCP server and see how it uses your tools.

  • Where to go next

    Developing with Neo4j MCP Tools covers using the official Neo4j MCP server, and Building GraphRAG Python MCP tools goes deeper on building custom MCP servers. Upcoming instructor-led sessions are listed on the workshops page.