SmartAPI Quick Start
This walkthrough starts with an API source and ends with generated scenarios ready for review. Generation produces drafts. Execute API requests during generation controls whether generation can call the API; Save scenario creates the independent AutoRunner copy for later execution.
Before you start
Confirm that:
- the correct RabbitQA workspace is active;
- your source is a supported API description, or you have a valid cURL request;
- secrets have been replaced with variables or safe placeholders; and
- you have permission to upload sources and start generation.
1. Open API Test Generation
Open SmartAPI → API Test Generation. The landing screen shows this workspace's totals — Sessions, Endpoints, and Generated scenarios — with View Sessions and Settings beside them, and recent sessions in a panel on the right.
Pick where the endpoints come from:
| Tab | Use it when |
|---|---|
| Use Inventory | The source has already been analyzed and published. Pick an inventory version and the sources within it |
| Upload File | You have a new API file. Drag it in, or browse to upload |
| Paste cURL | You only have requests. Paste one, choose Add cURL, and repeat for more |
Supported sources are OpenAPI, Postman, HAR, WSDL, GraphQL, Protobuf, and AsyncAPI — up to 5 files, 10 MB each. Uploaded files are analyzed automatically.
Use Inventory needs a published inventory version. If there is none yet, upload the source here or in File Management, then publish it from Inventory.
Select Continue to Discovery.
2. Review discovery and choose the scope
SmartAPI discovers the services, endpoints, and relationships in the source and summarizes them: endpoints discovered, services mapped, relationships detected, and the sources in scope.
Check that the service names and paths belong to your source, then select the endpoints to cover. At least one endpoint is required — until you pick one, the page says so instead of letting you continue.
Select Continue to Test Design.
3. Design the tests
Choose how discovery becomes test definitions.
| Method | What it does |
|---|---|
| AI Prompt Design | Describe the tests you want in natural language |
| Templates | Start from a Sanity, Smoke, or Regression suite |
| Import from Test Repository | Reuse existing test cases — one scenario is prepared per selected case |
The templates differ in breadth, and each shows how many scenarios it plans:
| Template | Coverage |
|---|---|
| Sanity | A focused check of the essential API behavior |
| Smoke | The critical paths and main operations |
| Regression | Positive, negative, boundary, authorization, and related-operation workflows |
Do not put passwords, tokens, or private customer data in the prompt.
Select Generate Tests.
Execute API requests during generation is a workspace default you can change per session. When it is on, generation sends real requests, which may create, change, or delete data. Check it before generating against anything shared.
4. Follow the session
Generation opens a session and reports its progress — selecting scenario workflows, generating scenarios, then writing the automation for each one. Scenarios appear as they are produced.
You can leave and come back: Sessions lists generation runs with their status, duration, and start method, searchable by name, file, or ID.
5. Review and save
Open a scenario and check its steps, requests, response checks, captured values, variables, and scripts before you keep it. Correct contract assumptions here rather than after transfer.
Save scenario writes it to an AutoRunner folder you choose. Each save creates a new, independent scenario, and the generated candidate stays available for further edits.
Saved scenarios are listed under Scenarios and are available in AutoRunner.
Next steps
- Inventory — Review endpoints and relationships, and publish the version generation uses
- File Management — Keep and version the sources you reuse
- Correct Use and Best Practices — The review gates that protect shared workspaces
