Neo4j

Neo4j is the leading native graph database, storing nodes and relationships directly on disk rather than reconstructing them with joins at query time. It runs self-managed as GPLv3 Community Edition or commercially licensed Enterprise Edition, and as the fully managed Aura cloud service, and is widely used for fraud detection, knowledge graphs and GraphRAG.

Neo4j is the market-leading native graph database and the company that popularised the property-graph model and the Cypher query language. Rather than storing rows and reconstructing relationships with joins at query time, Neo4j persists nodes, relationships and properties directly, so traversals that would require repeated self-joins in a relational database resolve at constant cost per hop — the reason it dominates fraud rings, identity graphs, supply-chain dependency analysis, recommendation engines and, more recently, GraphRAG pipelines that ground LLM answers in an explicit knowledge graph. The product ships in two lineages. Self-managed: Community Edition is GPLv3 open source, the core engine written in Java and built with Maven, currently on the calendar-versioned 2026.06 release with 17,000 GitHub stars, while Enterprise Edition adds closed-source clustering, role-based access control, hot backups and multi-database under a commercial licence. Managed: Aura spans AuraDB Free, Professional, Business Critical and Virtual Dedicated Cloud, plus a separate Aura Graph Analytics service for graph algorithms and data science. The wider ecosystem includes the Graph Data Science library, the APOC procedure library, a Spark connector, a Kafka connector, a GraphQL layer, an official MCP server and a graphrag-python package. Neo4j, Inc. was founded in 2007, is headquartered in San Mateo, California and is led by co-founder Emil Eifrem. Its June 2021 Series F raised $325 million led by Eurazeo with GV, DTCP, Lightrock, One Peak, Creandum and Greenbridge — at the time the largest round ever raised by a private database company — valuing it above $2 billion; that announcement claimed more than 800 enterprise customers and usage by over 75% of the Fortune 100, naming Adobe, AstraZeneca, eBay, Levi Strauss, UBS, Volvo Cars and Walmart. Third-party trackers put 2025 ARR near $200 million against roughly 1,000 employees and $535 million raised in total.

Category

Data & Analytics

Pricing

Subscription, Usage-based, Freemium, Contact for pricing

Target Market

CIOs, CTOs, Data Scientists, Enterprise Developers, Data Platform Engineers

Deployment

Cloud-first, Self-hosted, Open-source, Hybrid, Multi-cloud

Founded

2007

Headquarters

San Mateo, California, United States

Team Size

500+

Customers

800+ enterprise customers and 75%+ of the Fortune 100, as claimed in the June 2021 Series F announcement

• ✓
  Native property-graph storage

  Nodes, relationships and properties are stored directly on disk, so multi-hop traversals stay fast as data grows.

• ✓
  Cypher query language

  A declarative pattern-matching language that expresses relationship queries far more compactly than recursive SQL joins.

• ✓
  Graph Data Science library

  Ships production graph algorithms — centrality, community detection, similarity, node embeddings — for analytics and machine-learning pipelines.

• ✓
  Aura managed cloud

  Fully managed instances across AuraDB Free, Professional, Business Critical and Virtual Dedicated Cloud on the major hyperscalers.

• ✓
  Enterprise clustering and RBAC

  Enterprise Edition adds causal clustering, role-based access control, hot backups and multi-database isolation under a commercial licence.

• ✓
  GraphRAG and LLM tooling

  An official neo4j-graphrag-python package and an official MCP server let teams ground LLM answers in a queryable graph.

• ✓
  Broad integration surface

  Official Spark and Kafka connectors, a GraphQL layer, CDC client, APOC procedures and Docker and Helm deployment artifacts.

• •
  Fraud ring detection

  Surface rings of linked accounts, devices and payment instruments that row-based systems cannot connect in real time.

• •
  Knowledge graphs for GraphRAG

  Ground retrieval-augmented generation in an explicit graph so LLM answers cite traceable entities and relationships.

• •
  Identity and access path analysis

  Map who can reach what across systems to find privilege escalation paths and orphaned entitlements.

• •
  Supply-chain dependency mapping

  Trace multi-tier supplier and component dependencies to quantify blast radius when a single node fails.

• •
  Real-time recommendations

  Traverse user, product and behaviour relationships at request time instead of precomputing every candidate offline.

Aura is metered per GB of RAM per month — $65/GB on Professional and $146/GB on Business Critical, whose 2 GB floor puts it at about $292/month — so 100 GB on Professional runs roughly $6,500/month. There is no serverless tier, so you pay for provisioned capacity regardless of usage, and undersized memory silently degrades performance until you upgrade. Self-managed Enterprise licensing is quoted per core and never published; independent buyer data from Vendr puts the median Neo4j contract at $277,500 a year across a $66,400 to $912,000 range, with roughly 20% typically negotiated off list.

This page was written from 7 sources, 7 on domains other than neo4j.com.

1. 1.github.com — neo4j
2. 2.api.github.com — repos
3. 3.prnewswire.com — neo4j announces 325 million series f investment the largest
4. 4.toolradar.com — pricing
5. 5.vendr.com — neo4j
6. 6.peerspot.com — neo4j reviews
7. 7.getlatka.com — neo4j.com