AI Studio

Generate, manage, and review AI-powered test scenarios through a full-screen workspace with recording-based generation, Swagger/OpenAPI import, and an intelligent scenario tree.

What is AI Studio?

AI Studio is AutoRunner's dedicated interface for AI-assisted scenario management. It provides a full-screen workspace where you can generate test scenarios from user interaction recordings, import and generate scenarios from Swagger/OpenAPI specifications, organize scenarios in a tree structure, and review the output of AI generation sessions. AI Studio is accessible from the main AutoRunner navigation under AI Studio.

Interface Layout

AI Studio uses a two-panel layout:

• Left sidebar -- Tabbed navigation with scenario tree and management tools.
• Main content area -- Displays the active tab's content, including scenario details, generation workflows, and history.

Sidebar Tabs

Tab	Purpose
All Scenarios	Complete scenario tree with folder hierarchy and status indicators
Executable	Scenarios ready for execution (fully generated and reviewed)
Pending	Scenarios awaiting review or completion
Running	Scenarios currently being generated by the AI engine
Settings	AI model selection and prompt configuration
History	Generation history with session details and outcomes

Scenario Tree

The scenario tree displays all scenarios in a hierarchical folder structure. Each node shows:

• Scenario name and type icon
• Execution status indicator (passed, failed, pending, not run)
• Flaky badge -- Appears on scenarios with inconsistent pass/fail results across recent runs

Use drag-and-drop to reorganize scenarios between folders. Right-click a scenario or folder to access context actions including run, import, and delete.

Mass Operations

Select multiple scenarios in the tree to perform batch actions:

• Run Selected -- Execute all selected scenarios immediately.
• Import Selected -- Import selected scenarios into a test suite.
• Delete Selected -- Remove selected scenarios with confirmation.

Recording-Based Generation

Generate test scenarios by recording user interactions with your application. The AI analyzes the recording and produces structured BDD scenarios.

Workflow

1. Click New Recording in the AI Studio toolbar.
2. A browser session opens connected to your target application.
3. Perform the user workflow you want to test. AutoRunner captures every click, input, navigation, and assertion.
4. Click Stop Recording when finished.
5. The AI engine processes the recording and generates one or more test scenarios.
6. Review the generated scenarios in the Pending tab.
7. For each scenario, click Accept to add it to your scenario tree or Reject to discard it.

Recording Types

Recordings are stored with scenarioType=RECORDING and can be replayed or re-processed at any time. Each recording can produce multiple test scenarios covering different paths through the recorded workflow.

Swagger/OpenAPI Generation

Generate API test scenarios automatically from a Swagger or OpenAPI specification file. This is particularly useful for service-type projects that need broad API coverage quickly.

Workflow

1. Click Generate from Swagger in the AI Studio toolbar.
2. Upload your Swagger/OpenAPI specification file (JSON or YAML format).
3. A generation session is created displaying the detected endpoint count.
4. Configure generation options:

Option	Description
Scenario Count	Number of scenarios to generate per endpoint (1 to 20)
AI Provider	Select the AI model for generation (GPT-4, Claude, or Gemini)

5. Click Generate to start the asynchronous generation process.

Generation Stages

The generation process progresses through three stages, each tracked with a progress indicator:

Stage	Description
Preprocessing	Parsing the spec file, extracting endpoints, and building context
Generation	AI model generates test scenarios for each endpoint
Assembly	Scenarios are structured, validated, and organized into folders

The interface polls for status updates every 3 seconds. If generation does not complete within 10 minutes, the session times out and can be retried.

Generation Timeout

Large API specifications with many endpoints may approach the 10-minute timeout. If this happens, consider splitting your spec file or reducing the scenario count per endpoint.

Export Options

After generation completes, export the scenarios:

• Run Directly -- Add scenarios to the scenario tree and execute immediately.
• Export as Postman Collection -- Download a .json file compatible with Postman for external use.

AI Provider Selection

Configure which AI model powers scenario generation from the Settings tab:

Provider	Model	Best For
OpenAI	GPT-4	General-purpose scenario generation with strong reasoning
Anthropic	Claude	Complex workflows requiring detailed analysis
Google	Gemini	Fast generation for large specification files

You can also configure custom prompts in the Settings tab to guide the AI's generation behavior, such as emphasizing edge cases, security testing, or specific business rules.

Generation History

The History tab provides a complete record of all AI generation sessions:

• Session timestamp and duration
• AI provider and model used
• Number of scenarios generated
• Stage durations (preprocessing, generation, assembly)
• Final status (completed, failed, timed out)

From the history view, you can:

• Retry a failed or timed-out session with the same configuration.
• Accept or Reject individual scenarios from a completed session.
• Review the original input (recording or spec file) that produced the session.

Import from Repository

The Import from Repository view allows you to pull existing test files from a connected Git repository into AI Studio.

1. Select the target repository and branch.
2. Browse the file tree to locate test files.
3. Select files to import.
4. AI Studio parses the files and adds them to the scenario tree.

This is useful for teams migrating existing test automation assets into AutoRunner.

Device Generation

For mobile testing projects, the Generate for Device dialog creates device-specific scenario variants:

1. Select a base scenario from the tree.
2. Click Generate for Device.
3. Choose target device profiles (iOS, Android, screen sizes).
4. The AI adapts the scenario for the selected device contexts, accounting for platform-specific interactions.

Next Steps

• BDD Scenarios -- Write and manage test scenarios manually
• Editor -- Edit scenario code in the built-in Monaco editor
• Test Runner -- Execute generated scenarios
• Reports -- Analyze test results from AI-generated scenarios