Easy to Use

SQLFlow models stream-processing as SQL queries using the DuckDB SQL dialect. Express your entire stream processing pipeline—ingestion, transformation, and enrichment—as a single SQL statement and configuration file.

High Performance

Process 10's of thousands of events per second on a single machine with low memory overhead, using Python, DuckDB, Arrow and Confluent Python Client.

Powered by DuckDB

Tap into the DuckDB ecosystem of tools and libraries to build your stream processing applications. SQLFlow support parquet, csv, json and iceberg. Read data from Kafka.