About ZenML
ZenML is an open source framework for building machine learning and AI pipelines that can move between environments. The company also sells hosted AI engineering agents that work on top of ZenML and its open source sibling, Kitaru.
What it does
ZenML lets you write ML pipelines in Python and run them on different stacks without rewriting them. The company's hosted agents import traces from tools such as Langfuse, LangSmith and Braintrust, read your agent code from GitHub and look for gaps in your evals. They can test a fix as an experiment on past cases and report how the results change. Team members can ask these agents questions, and they can also work on tasks overnight.
Key features
- Write ML pipelines in Python that move between environments
- Import agent traces from Langfuse, LangSmith and Braintrust
- Read agent code from GitHub to find evaluation gaps
- Test prompt fixes as experiments on past cases
- Hosted AI engineering agents for debugging and evals
Who it's for
ZenML suits ML engineers and data teams who want portable pipelines. The hosted agents are aimed at teams that run AI agents in production and need help with evals and debugging.
Pricing and good to know
The site offers a free start and an open source install path. The hosted agents are a separate product with their own plans, . Some features are marked as coming next, so check what is available today before you plan around them.
Highlights
- Write ML pipelines in Python that move between environments
- Import agent traces from Langfuse, LangSmith and Braintrust
- Read agent code from GitHub to find evaluation gaps
- Test prompt fixes as experiments on past cases
- Hosted AI engineering agents for debugging and evals
