Skip to main content

Welcome

FlowLog is a composable Datalog engine that compiles programs into efficient and scalable Differential Dataflow executables. FlowLog is built for workloads that need to keep results fresh as data changes, without giving up on performance stability or operability at scale.

  • Incremental by design. FlowLog maintains results incrementally instead of recomputing from scratch, so updates are fast and efficient even over large fact sets.
  • Low latency with predictable performance. FlowLog takes care of join ordering and physical planning for you. You only write declarative logic rules while the compiler produces robust plans with stable, low-latency behavior—rather than brittle, hand-tuned queries.
  • Scales up and out. The same FlowLog program can run on a single powerful machine or across a cluster. The generated executables use Timely Dataflow for parallel and distributed execution.