Open-source tool to simplify ML model deployment_

  • Save your ML model with a Python call.

    Stick to your training workflow.

  • Model metadata is captured automatically.

    Use a human-readable YAML format for any ML framework.

  • Deploy models anywhere you want.

    Switch between deployment platforms with a single command.

  • Make Git a Model Registry

    MLEM is a core building block for Git-native ML model registries, combined with other tools like GTO or DVC.

We’re onGitHub
  • Tensorflow logo
  • PyTorch logo
  • dmlc xgboost logo
  • scikit learn logo
  • Light GBM logo
  • Keras logo
  • Catboost logo


  • Git-native ML model development. Enable GitFlow and other software engineering best practices.
  • Automatically detect ML framework, Python requirements, and data schema.
  • Seamlessly integrating to your stack thanks to Unix philosophy: one tool solves one problem very well.