function and several new data types to improve interoperability. Architecture & Design
To speed up similarity searches, create an index on the embedding column.
MATCH (d:Document) CALL d.embedding =~ [0.1, 0.2, ..., 0.n] // Your query vector RETURN d.content ORDER BY d._similarity_score DESC LIMIT 5; kuzu v0 120
The announcement came without much warning. In early October 2025, the main GitHub repository for Kùzu was , and a brief note appeared stating that "Kuzu is working on something new". The documentation and blog posts were also moved off the main website to GitHub.
Kùzu uses , the industry-standard query language for graphs. In v0.1.2.0, the engine has seen refinements in how it handles complex subqueries and aggregations. These improvements ensure that even the most deeply nested patterns are executed with minimal latency. 2. Storage Layer Optimizations function and several new data types to improve
Kùzu achieves state-of-the-art query processing speed through major structural innovations designed from the ground up for graph analytics:
MATCH (u:User) RETURN u.* // Returns all properties of the user node In early October 2025, the main GitHub repository
proves that the future of graph analytics is fast, efficient, and embedded. With its strong focus on analytical algorithms, easier migration paths from Neo4j, and expanded platform support, it is an essential tool for developers and data scientists handling large-scale, complex relationships. Whether you are building an AI-powered recommendation system or an offline graph analyst for mobile, Kuzu provides the performance needed to succeed. Looking to Get Started?
Crucial for network analysis, finding clusters, and determining network reliability.
If you want to see how these updates fit your specific project, tell me: What is your main stack built on?