Description
mlop is an open source platform for tracking, optimizing and collaborating on machine learning experiments. It logs parameters and gradients, tracks media such as images and audio, and sends alerts and notifications about runs. It tracks model performance and accuracy metrics and can flag problems in training. Reproducibility is supported by recording git status alongside experiments. The site describes it as compatible with the Weights and Biases API, so code that logs to that API can be pointed at mlop. The platform is backed by Y Combinator and includes compute and inference areas that the site marks as private beta or coming soon. Its pricing page offers a free tier for personal use or small teams with one seat, unlimited logging hours and 10 GB of storage, and a Pro plan for businesses scaling with AI. The source is on GitHub under mlop-ai/mlop.…
Preview
Pricing
- Free$0/mo
Highlights
- Experiment tracking for machine learning
- Parameter and gradient logging
- Media tracking and real-time alerts
- Reproducibility with git status
- Compatible with the Weights and Biases API
Ratings & Reviews
No ratings yet
Be the first to rate mlop.
People also visit

OptScale - MLOps & FinOps open source platform
MLOps and FinOps platform to run ML/AI experiments
mlop FAQ
What is mlop?
An open source platform for tracking machine learning experiments, including parameters, gradients and media.
Is it compatible with Weights and Biases?
The site says it has full Weights and Biases API compatibility.
Is there a free plan?
Yes. The free tier is for personal use or small teams, with one seat and 10 GB of storage.
Where is the code?
On GitHub at mlop-ai/mlop.
What are the best mlop alternatives?
Popular alternatives to mlop include MLRun.org, Weights & Biases and Nannyml. See all mlop alternatives




