The npm package kuzu@0.6.1-dev.36 is more notable because it was later flagged with two security vulnerabilities by Snyk:
import kuzu # Create or connect to a database persisted on disk db = kuzu.Database("./my_graph_db") connection = kuzu.Connection(db) # Define Node Schemas connection.execute("CREATE NODE TABLE User(id INT64, name STRING, age INT64, PRIMARY KEY (id))") connection.execute("CREATE NODE TABLE Hobby(name STRING, category STRING, PRIMARY KEY (name))") # Define Relationship Schema connection.execute("CREATE REL TABLE Follows(FROM User TO User)") connection.execute("CREATE REL TABLE Enjoys(FROM User TO Hobby)") Use code with caution. Step 3: Insert Data
The Kùzu project has moved far beyond those early versions. As of the latest information, the current stable release is (released on October 10, 2025). This version bundles four official extensions ( algo , fts , json , vector ), so you no longer need to install them separately.
Whether you are looking to enhance your AI stack with GraphRAG, untangle complex data relationships, or embed graph intelligence directly into your desktop or cloud applications, Kùzu v0.13.6 delivers a fast, stable, and serverless solution. Share public link kuzu v0 136
What is the of the dataset you intend to load? Share public link
: The system integrates with DuckDB , PostgreSQL , and formats like Parquet and Arrow . Current Status & Community Note
db = kuzu.Database("./my_graph_db") conn = kuzu.Connection(db) The npm package kuzu@0
The answer is an emphatic —especially if your workloads involve deep path traversals, nested property structures, or concurrent access patterns.
: Users can leverage the industry-standard Cypher query language. Vectorized Execution
This is an early pre‑release of the Rust client library, archived on docs.rs . Its description and feature list are identical to the current Kùzu documentation, indicating that even in its pre‑release state the core architecture was already defined. Developers experimenting with early Rust bindings would have used this version to test the embedded graph capabilities. This version bundles four official extensions ( algo
import kuzu # Initialize or open the database on disk db = kuzu.Database("./analytics_graph") conn = kuzu.Connection(db) # Create a Node Table for Users conn.execute("CREATE NODE TABLE User(id INT64, name STRING, age INT64, PRIMARY KEY (id))") # Create a Relationship Table for Follows conn.execute("CREATE REL TABLE Follows(FROM User TO User)") # Insert sample data using Cypher conn.execute("CREATE (:User id: 1, name: 'Alice', age: 30)") conn.execute("CREATE (:User id: 2, name: 'Bob', age: 25)") conn.execute("CREATE (:User id: 3, name: 'Charlie', age: 35)") # Establish relationships conn.execute("MATCH (a:User id: 1), (b:User id: 2) CREATE (a)-[:Follows]->(b)") conn.execute("MATCH (b:User id: 2), (c:User id: 3) CREATE (b)-[:Follows]->(c)") # Run an analytical 2-hop traversal query result = conn.execute( "MATCH (a:User)-[:Follows]->(b:User)-[:Follows]->(c:User) " "RETURN a.name AS Starter, c.name AS Target" ) while result.has_next(): row = result.get_next() print(f"row[0] is connected to row[1] via a 2-hop path.") Use code with caution. Interoperating with Pandas and Arrow
The release of brings a suite of performance optimizations, memory management updates, and storage enhancements. This version solidifies Kùzu's position as the go-to embedded graph database for developers who need Neo4j-like graph capabilities with the lightweight, serverless deployment model of SQLite or DuckDB.
Kùzu is an embedded, file-based graph database management system written in C++. Unlike traditional client-server databases (such as Neo4j), Kùzu operates directly inside your application process, eliminating network overhead. Think of it as the "SQLite for graph data." Core Architecture and Features