New experiment feature
We’re introducing a new experiment feature, now available to all users. This release is an early step to see how the feature works in practice and what we should improve next.
What’s included
Experiment creation
- Create experiments through a dedicated experiment workflow
- Experiments serve as the top-level container for:
- Trial definitions
- Parameters and bounds
- Results and visualizations
Experiments overview
- Centralized view of all experiments
- Clear separation between experiments and other platform entities
- Designed to support iteration and comparison across multiple experiments over time
Trial setup (Design of Experiments)
- Configure trials within an experiment using a design-of-experiments (DOE) approach
- Define input parameters and value ranges per trial
- Supports structured experimentation workflows rather than ad-hoc runs
- Trial configuration is kept inline to minimize context switching
Visualization
- Early-stage visualizations for experiment results
- Designed to help validate experiment behavior and outcomes
For more details on how this experiment works and what to expect, see the documentation:
Learn more about experiments
We especially encourage you to try it out using our Hartmann3 tutorial in the docs, which walks through a practical, end‑to‑end example.
Notes
- Access to a new capability that is still under active development
- Improvements may be rolled out iteratively during the experiment
- We’ll use usage data and user feedback to decide on the final scope and direction