Kuzu V0 136 • Trusted Source

: Enhanced performance for scanning JSON files during data ingestion. Database Architecture

It includes built-in HNSW vector indices and full-text search, making it a strong choice for GraphRAG and agent-based AI workflows.

You can install the latest v0.13.6 release directly via pip : pip install kuzu==0.13.6 Use code with caution. Basic Usage Example

Kùzu is not a toy database; it's a performant system packed with advanced features. kuzu v0 136

: A new mechanism to reclaim space automatically as you update or delete data in the database. Recursive Query Performance : Significant speed improvements for recursive queries, which are essential for deep graph traversals. JSON Scanning

npm install kuzu@0.136.0

No exact match for “kuzu v0 136” exists in any public dataset. : Enhanced performance for scanning JSON files during

Graph databases have transitioned from niche tools for academic research into critical infrastructure for modern enterprises. Whether powering fraud detection networks, driving real-time recommendation engines, or serving as the knowledge retrieval backbone (GraphRAG) for Large Language Models (LLMs), the demand for fast, efficient graph data management has never been higher.

The most practical improvement in v0.1.36 is the overhaul of the COPY FROM statement.

At the core of v0.1.3.6 are substantial under-the-hood enhancements to Kùzu’s storage engine and buffer manager. Graph databases are notorious for random memory access patterns due to pointer-chasing across large topologies. Kùzu mitigates this using a custom memory manager that uses smart caching and columnar storage. Version v0.1.3.6 refines memory allocation strategies during complex OPTIONAL MATCH and data ingestion routines, preventing memory spikes and ensuring predictable performance on resource-constrained machines. Enhanced Data Import Layouts Basic Usage Example Kùzu is not a toy

(e.g., log file, label on hardware, configuration file, error message, document)

Among these innovations is . Originating from academic roots at the University of Waterloo, Kùzu bridges the gap between relational-style performance optimizations and the flexible data structures inherent to graph networks.

Native support for vector indices (HNSW) and Full-Text Search (FTS) , making it a strong choice for AI-driven applications like GraphRAG.

Kuzu uses as its query language, ensuring a low learning curve for those familiar with modern graph systems. It also boasts a columnar storage engine optimized for both transactional (OLTP) and analytical (OLAP) workloads.

Before diving into version 0.136, it is important to understand Kuzu’s core philosophy. Unlike client-server graph databases like Neo4j or JanusGraph, Kuzu is an . It runs directly within your application’s process (similar to SQLite but for graphs). This design eliminates network overhead, making it uniquely suited for in-memory analytics, ETL pipelines, and edge computing.