Know exactly which data produced the result.
Version, classify and reserve test data beside the cases and executions that use it. Pin each execution to the inputs it used and trace their history.
Get started →- Draft data
- Classification
- Published version
- Test execution
- Usage history
Repeatable inputs
Know which data produced the result.
A test says “use a customer with an overdue invoice”. The shared account has changed, the spreadsheet is out of date, and the next tester cannot reproduce the failure. Nexus keeps versioned data sets alongside the tests that use them.
Publish a stable version
Define typed columns and identifiable rows, edit a draft, then publish it. Published versions stay unchanged. Compare versions row by row and pin a test to a specific version or let it follow the latest.
Reserve the right row
Limit a data set to the environments where it is valid. Consumable rows are reserved for a run so two executions do not unknowingly use the same account or identifier.
Reproduce the result
Pin an execution to the exact published version it used. Each selected row becomes an iteration, with usage history connecting the result back to its data.
Create, publish and use repeatable data.
Illustrative workflow
- Prepare. Create a synthetic “Overdue invoices” set with a row key, account reference and invoice balance. Restrict it to Staging.
- Publish. Publish version 2 after reviewing the rows and any scanner findings.
- Use. Pin the payment-reminder test to version 2. Run the selected rows as separate iterations within the execution.
- Investigate. When an iteration fails, follow its data usage to the version used. Publishing version 3 later does not change version 2.
This example describes a workflow, not customer results.
Build synthetic drafts with repeatable generators, import rows, or reserve and release data through the CI API. AI-assisted suggestions remain drafts for review before anything is published.
Make data handling a deliberate choice.
Classify each set as synthetic, masked or sensitive. Nexus flags likely personal data on import and edit, including New Zealand identifiers and email addresses outside reserved test domains. Sensitive values are masked, and reveals are logged.
Classification and scanning support review; they do not guarantee that data is anonymous or appropriate to use. Prefer synthetic inputs and review flagged values before publishing.