The Disappearance of Kuzu: A Cautionary Tale for AI and Knowledge Graph Development
author:: Volodymyr Pavlyshyn The open-source community recently experienced a significant loss that highlights the precarious nature of building on emerging technologies. Kuzu, a promising graph database that seemed perfectly positioned to revolutionize on-device AI memory systems, has been archived and is no longer supported — leaving developers who bet on its future scrambling for alternatives.
My LinkedIn Post
”The KuzuDB embedded graph database, open source under the MIT license, has been abandoned by its creator and sponsor Kùzu Inc, leaving its community pondering whether to fork or find an alternative.”
https://www.theregister.com/2025/10/14/kuzudb_abandoned/
It remains a great project that anyone can run locally and experiment with. Over three years of continuous development have added many features, including:
- Columnar, disk-based storage
- Embeddable, serverless integration into applications
- A flexible Property Graph Data Model and Cypher query language
- Import capabilities from a variety of columnar or relational stores
- A vectorized and factorized query processor
- Novel and high-performance join algorithms
- The Kuzu Explorer GUI
- Client APIs
- A RESTful API server
- Model Context Protocol (MCP) server for LLMs
- GraphRAG applications…
News YCombinator
https://news.ycombinator.com/item?id=45560036 I’ve been excited about Kuzu DB as a SQLite-style graph database. It looks like the devs are moving on to something else and no longer will support it, as of 10 October.
Their message reads, “Kuzu is working on something new! We will no longer be actively supporting KuzuDB. You can access the full archive of KuzuDB here: GitHub” https://github.com/kuzudb/kuzu
Oh too bad. Small fast embedded graph DBs are rare. Any good alternatives? There was a recent VLDB paper[1] demonstrating that the extension DuckPGQ[2] for DuckDB (an embedded database) offers competitive graph query performance compared to Neo4j and Umbra. No data on how it compares to KuzuDB. [1] https://vldb.org/cidrdb/papers/2023/p66-wolde.pdf [2] https://duckpgq.org/