mlop

mlop is an MLOps platform with a Weights & Biases-compatible API — an MLflow alternative for experiment tracking, model registry, and team collaboration. Self-hosted with W&B API compatibility.

Key features

Experiment tracking

• W&B-compatible API for seamless migration
• Real-time metric logging and visualization
• Hyperparameter tracking and comparison
• Rich media logging (images, audio, video)
• Experiment tagging and organization

Model registry

• Centralized model versioning
• Model lineage and provenance tracking
• A/B testing and model comparison
• Deployment-ready model artifacts
• Integration with ML frameworks

Collaboration

• Team workspaces and projects
• Shared experiments and notebooks
• Comments and annotations
• Role-based access control
• Usage analytics and quotas

Infrastructure

• Self-hosted with easy deployment
• Scalable backend with PostgreSQL
• REST API and CLI
• Integration with ML frameworks
• Webhook and notification support

At a glance

License	Apache-2.0
Stack	Python, TypeScript, PostgreSQL
Self-hosted	Yes — mlop OSS
Cloud	mlop.ai (managed)
API	W&B-compatible

Self-hosting

pip install mlop

mlop can be self-hosted with Docker or pip install. The cloud version provides managed infrastructure and team features.