kuzu
Highly scalable, extremely fast, easy-to-use embeddable graph database
What it is and what it does
Kuzu is an embedded graph database designed to run inside your application without a separate server. It stores data as property graphs (nodes and edges with attributes) and queries them using Cypher, a declarative graph query language. The database is optimized for analytical workloads on large graphs, offering features like full-text search, vector indices for similarity queries, and ACID transactions. It uses columnar storage and specialized join algorithms to achieve fast query execution across multiple CPU cores.
The package provides a Python binding to the Kuzu engine. You embed it directly in your application, point it to a local data directory, and issue Cypher queries programmatically. As of version 0.11.3, four commonly used extensions (algo, fts, json, vector) are pre-installed; earlier versions or additional extensions require setting up a local extension server. The project is now archived, meaning the maintainers have moved to a new initiative, but existing code using prior releases continues to work without modification.
Use it for:
- Build a recommendation engine by storing user-product relationships and querying graph patterns to find similar users or products.
- Analyze knowledge graphs or ontologies embedded in your application without deploying a separate database service.
- Perform full-text and vector similarity search on documents or embeddings stored as graph nodes.
- Run complex analytical queries on social networks, supply chains, or other highly connected datasets within a single process.
- Prototype graph-based features in a Python application before migrating to a production graph database.
Worth the install?
AI-flagged interpretation of the facts on this page — verify before relying
Kuzu is an embedded graph database that executes Cypher queries on property graphs, supporting full-text search and vector indices without requiring a separate server process.
Yes, if you need an embedded graph database for analytical workloads and can accept that the project is archived. The MIT license, zero runtime dependencies, and broad platform support make it low-friction to try. However, be aware that the maintainers are no longer actively developing new features—use it for stable, well-tested workloads, not for cutting-edge functionality or if you require ongoing vendor support.
Install
kuzu on PyPI
pip
pip install kuzuuv
uv add kuzupoetry
poetry add kuzuInstalling kuzu
Before you install
Medium install friction due to compiled wheels across multiple Python versions and platforms. The project is archived as of the latest release; prior versions remain usable but the maintainers are no longer actively developing new features.
License in practice
MIT license is permissive, allowing commercial and private use with minimal restrictions—suitable for most use cases without legal concern.
Quickstart
import kuzu
db = kuzu.Database(':memory:')
conn = db.get_connection()
result = conn.execute('MATCH (n) RETURN COUNT(n) AS count')
for row in result.get_as_df().itertuples():
print(row)
Requires a compatible Python version and matching platform wheel; no external runtime dependencies, but the database engine is compiled C++.
Verify before relying
- Whether the archived status affects long-term bug fixes or security patches for existing deployments
- Compatibility of pre-installed extensions (algo, fts, json, vector) with all use cases in v0.11.3
- Performance characteristics on specific graph sizes or query patterns
Package facts
| License | MIT (permissive) |
| Python support | not specified |
| Install friction | medium — platform-specific wheel |
| Runtime dependencies | none |
| Maintenance | abandoned — 308 days since the last release |
| Last repo commit | (repository archived) |
| First released | |
| Downloads | 436,393/month — #6,676 on PyPI (30-day window, as of 2026-08-14) |
| Known vulnerabilities | none known (OSV.dev, checked 2026-08-14) |
Evidence: kuzu-0.11.3-cp310-cp310-macosx_11_0_arm64.whl; kuzu-0.11.3-cp310-cp310-macosx_11_0_x86_64.whl; kuzu-0.11.3-cp310-cp310-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl; kuzu-0.11.3-cp310-cp310-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl; kuzu-0.11.3-cp310-cp310-win_amd64.whl; kuzu-0.11.3-cp311-cp311-macosx_11_0_arm64.whl; kuzu-0.11.3-cp311-cp311-macosx_11_0_x86_64.whl; kuzu-0.11.3-cp311-cp311-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl; kuzu-0.11.3-cp311-cp311-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl; kuzu-0.11.3-cp311-cp311-win_amd64.whl; kuzu-0.11.3-cp312-cp312-macosx_11_0_arm64.whl; kuzu-0.11.3-cp312-cp312-macosx_11_0_x86_64.whl; kuzu-0.11.3-cp312-cp312-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl; kuzu-0.11.3-cp312-cp312-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl; kuzu-0.11.3-cp312-cp312-win_amd64.whl; kuzu-0.11.3-cp313-cp313-macosx_11_0_arm64.whl; kuzu-0.11.3-cp313-cp313-macosx_11_0_x86_64.whl; kuzu-0.11.3-cp313-cp313-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl; kuzu-0.11.3-cp313-cp313-manylinux_2_27_x86_64.manylinux_2_28_x86_64.whl; kuzu-0.11.3-cp313-cp313t-manylinux_2_26_aarch64.manylinux_2_28_aarch64.whl
Keywords: graph, database
Tags
More Database packages
psycopg2-binary is a PostgreSQL database…
copyleft · top 1,000 on PyPI
redisPython client library for connecting to and…
permissive · top 1,000 on PyPI
ydbYDB Python SDK is the official client library…
permissive · top 1,000 on PyPI
snowflake-connector-pythonConnects Python applications to Snowflake data…
permissive · top 1,000 on PyPI
sqlparsesqlparse tokenizes SQL text into a tree of…
permissive · top 1,000 on PyPI
dbt-adaptersProvides base adapter protocols and shared…
permissive · top 1,000 on PyPI
real-ladybugEmbeddable graph database with Cypher query…
permissive · top 15,000 on PyPI
graph-retrieverCombines vector similarity search with graph…
permissive · top 15,000 on PyPI
FalkorDBFalkorDB Python client for querying and…
permissive · top 15,000 on PyPI
langchain-graph-retrieverRetrieves documents from a graph structure…
permissive · top 15,000 on PyPI
nucliadb-utilsProvides utility functions and data structures…
permissive · top 15,000 on PyPI
langchain-neo4jIntegrates Neo4j graph databases with…
permissive · top 15,000 on PyPI
neomodelAn Object Graph Mapper (OGM) for Neo4j that…
permissive · top 15,000 on PyPI
neo4j-graphragBuilds graph retrieval-augmented generation…
permissive · top 15,000 on PyPI
duckdbDuckDB is an in-process SQL database engine…
permissive · top 1,000 on PyPI