Local-first, open-source MLOps platform for ML beginners and small teams.
pip install zebraops
mlops init-project
mlops up
mlops doctor
mlops init my_model
mlops ingest my_dataset /path/to/data.csv
mlops train my_model my_datasetmlopsCLI with full lifecycle: init, ingest, train, eval, promote, serve, monitor- Contract-first model specs and immutable dataset manifests
- MLflow tracking, model registry, and promotion gates
- Prefect orchestration for reproducible pipelines
- FastAPI serving with Prometheus metrics
- Evidently drift monitoring with retrain recommendations
- Docker Compose local devstack (Postgres, MinIO, Grafana, Prometheus)
- Profile adapters for Vast, SageMaker, Vertex (extensible)
Apache License 2.0. See LICENSE.