github.com/MaterializeInc/materialize ↗
The live data layer for apps and AI agents. Create up-to-the-second views into your business, just using SQL
Open this visualization on its own page →
Contributors
46
Lines of Code
28,131
From
2017-12-05
To
2021-02-13
About MaterializeInc/materialize
Materialize is a real-time data integration platform written in Rust that creates and continuously updates consistent views of transactional data using SQL. It works by converting SQL queries into efficient dataflows that react to data changes as they happen, allowing organizations to serve up-to-the-second data without the complexity of building custom pipelines. The platform ingests data from PostgreSQL, MySQL, Kafka, and webhook sources, then lets users define views and perform queries through a PostgreSQL-compatible SQL interface.
The system is designed for three primary use cases: query offload patterns that scale complex reads more efficiently than read replicas, integration hubs that extract and transform data from multiple sources into live views, and operational data meshes that deliver real-time data products across services. Notably, Materialize guarantees correct and strongly consistent answers with minimal latency rather than accepting approximate or eventually consistent results, even when joining data across multiple upstream systems. The platform supports comprehensive SQL features including complex multi-way joins, delta-joins that avoid intermediate state blowup, subquery decorrelation, arbitrary updates and deletes, and JSON operations.
Materialize is available as a fully managed cloud service with high availability and horizontal scalability, as well as self-managed Community and Enterprise editions. The community edition is free for deployments under 24 GiB of memory and 48 GiB of disk. The project is actively developed and welcomes external contributions, with a focus on expanding PostgreSQL feature compatibility.