Getting started
Overview
What the AI Generator does
Transform requirements into production-ready test cases using advanced AI models.
- Generate test cases from natural language requirements
- Choose from multiple AI models (Claude, GPT)
- Customize test count and coverage depth
- Save templates for consistent generation across projects
- Generate platform-specific tests (Web, Mobile, API, Accessibility, Performance)
AI-powered intelligence
The generator analyzes your requirements and automatically creates positive tests, negative tests, edge cases, and boundary conditions based on your selected coverage level.
Prerequisites
What you need before generating test cases.
- Active SynthQA account
- Clear requirements (quick entry or saved requirements)
- Understanding of your test coverage needs
Choose your approach
Generation Types
Regular vs. Cross-Platform
Two generation modes for different testing needs.
Regular Test Cases
Generate traditional test cases for specific requirements or features.
- • Focused on single feature/requirement
- • Quick setup and generation
- • Flexible requirements input
- • Save and reuse requirements
Cross-Platform Testing
Generate platform-optimized tests for Web, Mobile, API, and more.
- • Test across multiple platforms
- • Framework-specific test cases
- • Ensure cross-platform consistency
- • Separate suites per platform
| Feature | Regular Generation | Cross-Platform |
|---|---|---|
| Best for | Single feature testing | Multi-platform consistency |
| Output | One test suite | Suite per platform |
| Setup complexity | Simple | Moderate (platform selection) |
| Use templates | ✓ Supported | ✓ Supported |
Which should I use?
Use Regular for feature-specific testing. Use Cross-Platform when you need to verify that functionality works consistently across web, mobile apps, APIs, or accessibility standards.
Step-by-step
Regular Test Case Generation
Multi-device
Cross-Platform Testing
Cross-platform workflow
Generate platform-specific test suites in one operation.
Cross-platform testing generates separate test suites optimized for each selected platform and framework.
Web Application
Browser testing
Mobile App
iOS/Android
API/Backend
REST/GraphQL
Accessibility
WCAG compliance
Performance
Load testing
Efficiency
Using Templates
Template system
Save and reuse your preferred generation settings.
What templates save:
- AI model selection
- Test case count
- Coverage level
- Optional: edge cases and negative test preferences
Template categories:
Functional: Standard feature testing
Security: Auth, permissions, encryption
Performance: Load, stress, response times
Integration: API, third-party services
Regression: Existing feature validation
Accessibility: WCAG, screen readers
Options
AI Models
Choosing the right model
Different models offer different tradeoffs in speed, cost, and quality.
Claude Sonnet 4.5
RecommendedBest for: Most test generation tasks
Speed: Fast (10-15 seconds)
Quality: Excellent balance
Use when: You need reliable, thorough test cases quickly
Claude Haiku 4.5
FastestBest for: Quick iterations, drafts
Speed: Very fast (5-8 seconds)
Quality: Good for simple requirements
Use when: You need rapid generation for straightforward tests
Claude Opus 4.5
Maximum QualityBest for: Complex, critical features
Speed: Slower (20-30 seconds)
Quality: Highest depth and creativity
Use when: Testing complex business logic or security-critical flows
GPT Models (4o, 4o-mini, 5, 5.2)
Alternative models with different strengths. GPT-4o offers fast, reliable results; GPT-5 series provides cutting-edge reasoning for complex scenarios.
Model recommendations
Start with Claude Sonnet 4.5 for most use cases. Use Haiku for rapid iterations. Reserve Opus for your most critical or complex requirements.
Depth
Coverage Levels
Understanding coverage levels
Coverage determines how thorough and comprehensive your test cases will be.
Standard Coverage
Includes:
- Happy path scenarios
- Basic positive test cases
- Core functionality validation
- Common user workflows
Best for: Quick validation, smoke testing, initial feature checks
Comprehensive Coverage
DefaultIncludes:
- All standard coverage
- Edge cases and boundary conditions
- Negative test cases
- Input validation scenarios
- Error handling paths
Best for: Production features, regression testing, standard QA cycles
Exhaustive Coverage
Includes:
- All comprehensive coverage
- Extensive boundary testing
- Complex interaction scenarios
- State transition coverage
- Security and permission tests
- Performance edge cases
- Rare but critical scenarios
Best for: Critical features, security-sensitive flows, compliance requirements
Consider test maintenance
Higher coverage means more test cases to maintain. Balance thoroughness with practicality based on feature criticality.
Quality
Best Practices
Writing effective requirements
The quality of your test cases depends on the clarity of your requirements.
✅ Do
- Be specific about expected behavior
- Include validation rules and constraints
- Mention error states and messages
- Specify data types and formats
- List user actions step-by-step
- Include boundary values (min/max)
❌ Avoid
- Vague or ambiguous descriptions
- Missing validation criteria
- Assuming implicit behavior
- Overly technical jargon without context
- Single-line requirements
- Mixing multiple features in one requirement
Test case organization
Structure your generations for maintainability.
- One requirement or feature per generation
- Use descriptive titles that indicate scope
- Tag test cases with relevant labels after generation
- Group related generations by feature area
- Use projects to organize by sprint or release
Model selection strategy
Choose models based on requirement complexity.
Haiku 4.5
• Simple CRUD operations
• Basic form validation
• Straightforward UI flows
Sonnet 4.5
• Most features (80% of cases)
• Multi-step workflows
• Standard business logic
Opus 4.5
• Complex algorithms
• Security-critical features
• State management systems
Performance
Optimization Tips
FAQ
Frequently Asked Questions
Help
Support
Getting help
Resources and contact information for assistance.
Contact options:
- Email: support@synthqa.app
- Use the feedback button in the dashboard
- Check our knowledge base for common issues
When reporting generation issues, include:
- Your requirement description
- Selected model and settings
- What you expected vs. what was generated
- Generation ID (if available)
Last updated: January 2026 · Guide version: 1.0