Skip to content

Introduction

AntFlow is a platform that turns scientific and geospatial workflows into production-ready processing services.

Using a GitOps approach, workflows stored in Git repositories can be exposed through standard APIs and executed across Kubernetes, HPC, cloud, and on-premise infrastructures.

Why AntFlow?

Bringing scientific workflows into operational use is challenging. Beyond workflow development itself, teams often need to manage infrastructure, deployments, APIs, monitoring, traceability, interoperability, and data cataloging before a process can be reliably operated in production.

AntFlow provides these capabilities out of the box, allowing scientists, engineers, and platform teams to focus on building workflows rather than operating the platform that runs them.

Key features

  • Publish workflows directly from Git repositories and turn them into production services.
  • Expose workflows as standard APIs through OGC API - Processes endpoints.
  • Run across Kubernetes clusters, HPC systems, cloud environments, and local infrastructure.
  • Preserve execution lineage, parameters, logs, and outputs.
  • Catalog results for discovery and reproducibility.

Adoption and sponsorship

AntFlow is an open-source project, sustained by its adoption and sponsorship from leading European space organizations and technology companies, including CNES, ESA, and CS Group.

AntFlow adopters

AntFlow powers Earth Observation, Digital Twin, and scientific processing platforms including the CNES Digital Twin Factory, FloodCARE, STAC Planet, ESA IF Pathfinder (Urbi Verde), CNES Datalake (Geodes), OGC OSPD, GPOS, and CNES Streams.

See the Blog for publications and talks about AntFlow.

What's next

  • Try AntFlow: open the hosted demo. The User guide explains how to run workflows and inspect results.
  • Understand how it works: see the Architecture for the components and the path from a workflow to its results.
  • Deploy your own instance: follow Installation with Helm to install AntFlow on Kubernetes with Helm.
  • Operate an existing instance: use the Operator manual to find service, access, and execution configuration guides.