Kuzu

Embedded, Python-native graph database optimized for local AI research and data science workflows.

Overview

Kuzu is a lightweight, embeddable property graph database that runs directly in memory or as a Python library, eliminating server setup overhead. It offers a SQL-like query language with deep Python integration, making it ideal for exploratory data analysis, local LLM tooling, and academic AI research. The engine supports fast graph construction from heterogeneous sources and scales cleanly from notebook experiments to modest deployments.

Best for

• Zero-config local GraphRAG experimentation in Jupyter/VS Code
• Data science teams embedding graph logic directly into Python pipelines

Trade-offs

What XeroHack pre-wires

When you pick Kuzu in the interview, the engine emits these files into your scaffold:

• pip install kuzu + in-memory graph initialization
• Pandas DataFrame to graph node/edge conversion helpers