Skip to main content
Create and run experiments in the Braintrust UI without writing code. Build one from scratch with a dataset and scorers, or promote a playground run you’ve iterated on to an immutable snapshot.

​
Create from scratch

  1. Go to Experiments.
  2. Click + Experiment or use the empty state form.
  3. Select one or more tasks to evaluate: prompts, workflows, or remote evals and sandboxes.
  4. Choose or create a dataset:
    • Select existing dataset: Pick from datasets in your organization
    • Upload CSV or JSON: Import test cases from a file
    • Empty dataset: Create a blank dataset to populate manually later
  5. Optional: Click Filter under the dataset field to scope the experiment to a subset of records. Filter by input, expected, metadata, tags, or any custom column using SQL expressions. The experiment runs only against records matching the filter, and the filter is recorded on the experiment.
  6. Use the Scorers field to add scorers and classifiers to score and label outputs.
  7. Click Create to execute the experiment.
UI experiments run without a time limit on cloud and on self-hosted deployments running data plane v2.0 or later.

​
Promote from a playground

Playground runs are mutable — re-running overwrites previous results. When you’ve iterated to a configuration worth keeping, promote it to an experiment to capture an immutable snapshot:
  1. Run your playground.
  2. Select + Experiment.
  3. Name your experiment.
  4. Access it from the Experiments page.
Each playground task maps to its own experiment. Experiments created this way are comparable to any other experiment in your project.

​
Next steps