Curate from production

Turn real production spans and experiment results into dataset rows — recover prompt variables from traced calls, link back to the source log, or export experiment failures as a regression set.

The best test cases come from real traffic and real eval runs. Trodo can promote a production span or an experiment's rows into a dataset without hand-copying inputs and outputs.

From a trace span

  1. Open Agent Runs or Traces and find the run.
  2. Click into the span detail for the LLM call you want to capture.
  3. Click Add to dataset.
  4. Pick an existing dataset or type a new name.

Trodo turns the span into a row:

FieldSource
InputWhen the span carries a trodo.prompt.version_hash, Trodo resolves that exact prompt version and recovers the variable values the call was compiled with. Otherwise it stores parsed query / system / context fields.
Expected outputThe span's recorded output.
OriginA back-link to the source span — click Log in the grid to jump back.

Prompt-attributed spans recover richer variable bindings via trodo.prompt.version_hash. See Prompt traceability for how version hashes are stamped on spans.

Rows keep a source_trace_ref back-pointer. The Origin column in the dataset grid shows Log for trace-sourced rows.

Prerequisites

You need tracing wired up so production calls appear in the dashboard. For managed prompts, compile inside a tracked agent run so the span is stamped with the prompt version hash.

From an experiment

After running an experiment, open the experiment detail page and click Create dataset from experiment (database icon).

Configure:

OptionChoices
Dataset nameName for the new dataset (defaults to experiment name + "dataset").
Rows to includeAll rows · Failures only (error or composite < 0.70) · Passes only (composite ≥ 0.70)
Expected output sourceKeep original ground truth from the source dataset · Use experiment model output as expected output
ModelWhen the experiment ran multiple models, pick which model's output to use (defaults to first result per row).

When to use each mode

GoalFilterExpected source
Regression set from failuresFailures onlyKeep original ground truth
Promote model answer as new truthAll or passesUse experiment output
Full copy of experiment inputsAllKeep original ground truth

Experiment-exported rows show Exp in the Origin column — click to open the source experiment.

Compare the two sources

Trace spanExperiment
Best forA specific production failure with rich prompt contextBatch-exporting eval results
InputRecovered prompt variables or parsed span fieldsOriginal dataset input
ExpectedSpan outputOriginal ground truth or model output
Origin linkLog (span)Exp (experiment)

After curating

  1. Open the new dataset and review rows in the dashboard editor.
  2. Add tags (e.g. regression, prod-incident-2026-03) for filtering.
  3. Click Save, then optionally create a snapshot.
  4. Run an experiment against the curated set to verify a fix.

Next

On this page