Neo4j Freemium

Neo4j is a graph database that stores data as nodes and relationships and queries them with Cypher. It runs on-premise or as AuraDB in the cloud, and now ships vector indexes for combining graph traversal with embeddings.

What is Neo4j?

Neo4j started in 2007 and is the de facto graph database. Instead of tables and joins, you write (:Person)-[:BOUGHT]->(:Product) patterns. Traversals that would need five recursive joins in SQL run as one hop.

The 2025 releases (Neo4j 2025.x and Aura) added vector indexes plus the GraphRAG Python package, aimed at teams building retrieval over connected documents rather than isolated chunks.

Product Features

• Cypher query language — pattern matching designed for paths; shortest-path and variable-length traversals are built in.
• Vector index — store embeddings on nodes and combine similarity search with graph filters in one query.
• GraphRAG package — extract entities with an LLM, write them to the graph, retrieve subgraphs for context.
• AuraDB — managed cloud with automatic backups; Aura Free gives 200,000 nodes and 400,000 relationships.
• APOC and GDS — dozens of procedures plus a graph data science library for centrality, community detection and link prediction.
• Drivers — official support for Python, Java, JavaScript, .NET and Go.

Product Highlights

Accept the modelling cost and the payoff is real. A fraud ring, a supply chain or a recommendation graph becomes something a product manager can read as a diagram. Query plans show you exactly which hops fire.

Pricing: Aura Free at no cost; Aura Professional from $65 per month for 4 GB RAM and 8 GB storage. Enterprise and on-prem licensing is quoted per core and usually starts in five figures per year.

Honest downsides: writes are slower than Postgres for bulk inserts, the community edition lacks clustering, and hiring people who can model graphs well is harder than hiring SQL developers. If your data has no meaningful relationships, a relational database will be simpler and cheaper.

Who should use it (and who should not)

Good fit: fraud detection, master data management, recommendation engines, and GraphRAG projects where relationships carry the meaning. Bad fit: straightforward CRUD apps or analytics over flat event tables — that is Postgres or ClickHouse.

Use Cases

• GraphRAG over contracts where parties, clauses and dates are linked.
• Identity resolution across accounts that share devices or addresses.
• Recommendations built on co-purchase and co-view paths.
• Impact analysis in a supply chain with multi-tier suppliers.
• Knowledge graphs that sit behind an internal AI assistant.

Quick Summary

• Neo4j stores nodes and relationships and queries them with Cypher.
• Aura Free covers 200,000 nodes and 400,000 relationships.
• Aura Professional starts at $65 per month; Enterprise is quoted.
• Vector indexes and the GraphRAG package make it a common LLM retrieval backend.
• Bulk writes and graph modelling skill are the real costs.