PuppyGraph

PuppyGraph is a zero-ETL graph query engine that runs graph analytics directly against existing relational databases and data lakes, without requiring data to be extracted, transformed, or loaded into a separate graph database. It connects to data sources — including Snowflake, Databricks, BigQuery, Redshift, Apache Iceberg, Delta Lake, PostgreSQL, and MySQL — and exposes the underlying relational data as a property graph, queryable via Gremlin or openCypher. No data duplication or synchronisation pipeline is required. The query engine is distributed and columnar, designed for OLAP-style graph workloads at scale, supporting petabyte-range datasets. It supports both SQL and graph query languages against the same data simultaneously. It is aimed at organisations that want graph analytics on data already held in a data warehouse or lakehouse, without the operational overhead of deploying and maintaining a dedicated graph database. Primary use cases are fraud detection, network analysis, knowledge graph querying, and GraphRAG pipelines. It was founded in 2023 and remains at seed stage.