videoflow’s documentation

Videoflow is a Python framework for building distributed video and stream processing pipelines. You describe your pipeline once as a directed acyclic graph of producers, processors and consumers, and Videoflow runs it as a set of independent workers that communicate over a NATS JetStream message broker.

The same graph runs two ways with no code changes:

Locally

As one OS subprocess per node — for fast development and testing.

On Kubernetes

As one container per node — with per-node scaling, GPU scheduling, health probes and autoscaling for production.

Developer friendly

Even complex pipelines are defined in a small build_flow() factory of a few lines of code.

Reliable

At-least-once delivery with ack-after-process, per-message de-duplication, retries and a dead-letter queue, so a crash or a bad message never silently loses or double-emits data.

Easy to extend

Writing your own producers, processors and consumers is straightforward — sync or async, with an optional runtime context.

Language-agnostic

Components can be written in any language, shipped as container images, and dropped into a Python-authored graph by reference — the basis for a component marketplace.

Free and open source

Videoflow uses the MIT License.