AI Potluck
Contribute
Back to Gap Map Infrastructure / ML orchestration

MLRun

MLRun
open source / Overall score: 2.4

Open-source MLOps platform for building and running continuous machine-learning and generative-AI applications. Functions are declared in Python and executed as jobs on Kubernetes, with pipelines, a feature store, artifact and model tracking, and real-time serving graphs over the same objects. It is the open core beneath Iguazio's commercial platform.

Iguazio, MLRun's commercial steward, is now part of McKinsey.

Openness

5 high confidence
5.0
license
Apache-2.0(OSI)
source
public(mlrun/mlrun is the runnable platform, self-host included)
core features withheld
no — a 2,580-entry recursive tree carries one root LICENSE and no path matching enterprise, ee, commercial or proprietary

MLRun is Apache-2.0, with the whole platform, self-hosting included, in the open repository. Iguazio sells a managed offering alongside it, but MLRun's own project site describes the open framework itself as the product, and the repository's LICENSE covers the whole tree with no enterprise, ee or commercial path.

  • https://api.github.com/repos/mlrun/mlrun/git/trees/development?recursive=1 recorded 2026-09-15

    Full untruncated recursive tree of the default branch, 2,580 entries. One root LICENSE. No path matches enterprise, ee, commercial or proprietary. A tree lists paths; it is cited for that and for nothing about the vendor's offerings.

  • https://github.com/mlrun/mlrun recorded 2026-09-15

    Repository page for mlrun/mlrun: Apache-2.0 license, public and unarchived, described as an open source MLOps platform for building and managing continuous ML applications. Establishes the license and that the source is public.

  • https://www.mlrun.org/ recorded 2026-09-16

    MLRun's own project site: 'MLRun is an open-source AI orchestration framework for managing ML and generative AI applications across their lifecycle.' The vendor describes the open framework as the product rather than as a reduced edition of a paid one.

Adoption

2 high confidence
2.0

Adoption is measured by PyPI downloads of mlrun, its declared distribution channel.

Capability

3 medium confidence
3.0

MLRun treats runs, models and artifacts as first-class objects it schedules and tracks: it decides when a function runs and on which cluster. The Kubernetes scheduler chooses the actual machine, which keeps MLRun level with Kubeflow Pipelines rather than higher.

  • https://github.com/mlrun/mlrun recorded 2026-09-15

    Repository page and README for mlrun/mlrun describing functions executed on Kubernetes, pipelines, the feature store and model tracking.

Verified 2026-09-15