Correct Use and Best Practices

SmartAPI accelerates API test design, but generated content remains a reviewable draft. Apply these gates before sharing or executing the result.

Use the right workflow

Need	Workflow
Fast generation from a known source	API Test Generation, scoped to a published inventory version
Separate discovery and design review	API Test Generation
Reuse and version API contracts	File Management
Return to a saved run	API Test Sessions
Execute reviewed scenarios	Save the scenario to AutoRunner, then run it in AutoRunner

Verify provenance

Before editing or saving a scenario, record or compare:

• active RabbitQA workspace;
• managed file and version or cURL origin;
• SmartAPI session ID;
• recognizable endpoint paths;
• service and endpoint counts; and
• the selected template or design prompt.

If discovery or session detail shows unrelated paths, check the selected source and inventory version before editing.

Protect credentials and data

• Replace authorization headers, cookies, API keys, passwords, and client secrets with variables.
• Use synthetic examples instead of real customer, employee, or company data.
• Inspect uploaded OpenAPI examples, Postman files, HAR files, and cURL commands for hidden secrets.
• Keep runtime secrets in the authorized AutoRunner execution configuration.
• Do not paste confidential data into AI prompts or scripts.

Review generated assets

Check every request method and path against the source contract. Validate bodies, parameters, headers, assertions, captures, and variable ordering. Review positive and negative classifications and remove duplicate or unsafe scenarios.

Use Inventory to inspect endpoint relationships and their evidence. Use the source contract to validate generated automation, and the AutoRunner report to establish execution results.

Treat destructive actions carefully

• Retain a session when its generation history may be needed.
• Confirm file-use and version history before deleting a managed source.
• Review the replacement scope before regenerating a scenario or its later steps.
• Confirm the session ID before permanent deletion.
• Never run generated destructive requests against production without separate authorization and purpose-built safeguards.

Saving gate

Save only after source identity and the scenario automation have been reviewed. Saving creates an independent AutoRunner copy; it does not start a run. In AutoRunner, check the saved scenario, environment, run profile, test plan, and final execution report.

Troubleshooting

Problem	What to check
Generate is disabled	Select at least one endpoint, then complete the prompt or choose a template or repository case.
A managed file is missing	Confirm active workspace, search term, type filter, and source version.
Discovery shows unrelated paths	Stop; compare session/source identity and do not use reference data as a result.
Immediate and saved counts differ	Use the persisted session ID and session detail as the durable run reference after verification.
Saving is unavailable	Confirm scenario generation has finished, your update permission, and the destination folder.
Saved scenarios do not run	Configure AutoRunner environment values, secrets, scenario order, run profile, and test plan.
File analysis failed	Check the source format and content, then use Re-Analyze in File Management.