What to expect at NODES 2026
Live sessions
around the clock
Get ready for 24 hours of live sessions, lightning talks, and more – in every major time zone across the globe.
Content to match
your interests
Gain expert insights on main topics: multi-agent systems, modern applications, and deep context with graphs.
Learning about the
latest and greatest
Find fresh ideas and inspiration when you hear about AI engineering, data intelligence, tech stack integrations, and more.
Meaningful
connections
Connect with speakers during Q&As, exchange ideas, and find new collaboration opportunities in the community.
Keynote Fireside Chat with Andrew Ng & Emil Eifrem
Building complex agentic systems requires new approaches to engineering, context, and data. In this fireside chat, Dr. Andrew Ng, a global leader in AI innovation and education, joins Emil Eifrem, CEO and Co-Founder of Neo4j, to explore what’s next for AI builders.
Andrew Ng
Dr. Andrew Ng is a pioneer in machine learning and online education who has taught AI to over 10 million people. He founded the Google Brain team, served as VP & Chief Scientist at Baidu, and directed the Stanford AI Lab. In 2023, he was named to the Time100 AI list of the most influential people in AI. Read his bio.
Emil Eifrem
Emil Eifrem sketched what today is known as the property graph model on a flight to Mumbai and has devoted his professional life to building, innovating, and evangelizing graph databases and graph analytics. Read his bio.
Bring your friends to NODES 2026 for a chance to win $200 worth of cool prizes
Top 30 participants with the most friends attending will each receive $200 worth of cool prizes from the Neo4j Swag Shop. We’re tracking attendance, so be sure your friends register AND attend! Check out the contest rules for more information.
Road to NODES workshops
Get ready for NODES 26! All workshops are 2 hours, free and virtual. Register for workshops here.
Build a lightweight knowledge layer over your existing documents and SQL data using three shapes in Neo4j: a Table of Contents (Trees) for document navigation, Themes (Communities) that surface unnamed patterns, and Connections (a Semantic Layer) linking structured entities and records.
October 22 | 17:00 BST / 18:00 CEST / 9:00am PDT / 12:00pm EDT
Two technical sentences:
Run Cypher queries directly against your cloud data warehouse — Snowflake, Databricks, or Google BigQuery — using Neo4j Virtual Graph, mapping relational data into a graph model in place without moving or duplicating it.
October 29 | 17:00 GMT / 18:00 CET / 10:00am PDT / 1:00pm EDT
Build a full graph-based agent memory stack in Neo4j — hierarchical context graphs, retrieval and answer playbooks, and dynamic playbook composition — and compare how predefined workflows versus fully autonomous agents interact with that memory differently.
November 5 | 17:00 GMT / 18:00 CET / 9:00am PST / 12:00pm EST
Work hands-on against a running Neo4j Agent Memory Service (NAMS) instance to build a memory layer holding three kinds of memory in one graph: short-term conversation history, long-term entities modeled with POLE+O, and reasoning memory that captures decision traces and tool-call provenance as first-class nodes.
Here’s how the global graph
community showed up this past year
22,453
Registrations
150
Countries
4,434
Companies
Frequently asked questions
NODES is Neo4j’s annual online conference for AI practitioners building context pipelines for AI systems. It’s free, runs for 24 hours straight on November 12, with 100+ live sessions across time zones. Developers, data teams, and AI practitioners are sharing their current projects, practical approaches, and lessons learned. With 22,000+ enthusiasts from 150 countries over the past year, it’s the largest gathering of the graph community.
Bring friends! The 30 people who get the most friends to register and attend win $200 worth of Neo4j swag.
AI engineers, developers, data scientists, architects, and non-technical folks looking to build better context and workflows for agentic systems. You don’t need to be a graph expert. There are sessions for beginners and for teams already running this in production. Topics fall into three areas:
- AI engineering: building with generative AI, GraphRAG, and agentic workflows.
- Modern applications: using graphs with whatever language and stack you already use.
- Data intelligence: using knowledge graphs to get more out of your data.
- How do I reduce hallucinations and trace every answer back to its source?
- How do I give my agents memory that lasts beyond one conversation?
- How do I get better results from queries that span multiple data sources and require multi-hop reasoning?
- How do I create the right ontology or knowledge graph for my use case in minutes?
- How do I help my agents navigate disconnected data and complex workflows?
- How do I turn unstructured data into meaningful context for my AI agents?
- When do I need a knowledge graph, and when are vectors enough?
- How do I get a prototype into production?
- What are other teams building, and what did they learn?
Register for free today at neo4j.com/nodes
GraphRAG: When you ask an AI a question about your company’s information, it first has to find the right pieces. Most systems look for text that sounds similar to your question using vector-only and RAG. GraphRAG adds a map of those connections, then traverses graph relationships (entity links, dependencies, time-based paths) to collect the full context. You get better accuracy and traceable answers with fewer tokens sent to the LLM.
Agentic memory and context: AI agents need to remember past interactions, connect facts, and use the right context to make decisions. A graph-based memory layer for episodic, semantic, and procedural memory helps agents recall relevant information, update what they know, and share knowledge across agents. With AI integrations, you can build with the frameworks you already use and keep your data in your existing platforms.
Ontology: Your business’s rules for what things exist and how they relate. Most companies have this knowledge scattered across documents, spreadsheets, and in many locations. You can bring your data in where it is, build a graph so you can see and change it, and then use it to organize your data so your AI follows the same rules your business does. You don’t need to get it perfect up front. You can start small and grow it.
Yes. The free Road to NODES workshops are two hours each and run on Thursdays through November 5.
No. You can start free with Neo4j AuraDB Free in the cloud, on desktop, or run it locally in Docker.