Enhance AI testing accuracy by building a comprehensive knowledge base of your application. Stably’s AI uses this knowledge to understand your app’s functionality, user flows, and context for more accurate test creation and execution.
App Knowledge is completely optional. Stably’s AI can create and run tests effectively without any additional knowledge. However, providing app knowledge helps the AI generate more accurate tests and better understand your application’s specific context and workflows.

​
Overview

App Knowledge helps AI understand your application better and perform more accurate testing through multiple knowledge sources:
  • User Documentation: Product documentation and help articles
  • Session Recordings: Video recordings of user interactions
  • Test Execution Data: Screenshots and data from past test runs
  • Manual Knowledge: Custom documentation and context you provide

​
Knowledge Sources

​
User Documentation

Connect your existing documentation to provide AI with comprehensive understanding of your application’s features and workflows.
1

Navigate to User Docs

Access the App Knowledge section and select User Docs
2

Add Documentation URLs

Click Edit web links to add your documentation URLs
3

Configure Crawling

Specify which sections of your documentation to include in the knowledge base
4

Index Documentation

Stably automatically crawls and indexes your documentation daily, or you can manually trigger re-indexing
Supported Documentation Sources:
  • Product documentation sites
  • API documentation
  • User guides and tutorials
  • Knowledge base articles
  • Help center content

​
Documentation Processing

Stably automatically:
  • Extracts Text: Processes all text content from your documentation
  • Processes Images: Analyzes screenshots and diagrams
  • Creates Searchable Index: Makes content available for AI queries
  • Updates Regularly: Refreshes content daily or on-demand

​
Session Recordings

Upload video recordings of user sessions to help AI understand real user interactions and workflows.

​
Adding Session Recordings

1

Access Session Recordings

Navigate to App Knowledge > Session Recordings
2

Upload Video Files

Drag and drop video files or click to upload
  • Supported formats: .mp4, .mov, .mkv, .webm
  • Maximum file size: 1000MB
3

Process Recordings

Click Start Processing to analyze the uploaded recordings
4

Review Knowledge

Processed recordings become available as searchable knowledge for AI testing
Recording Sources:
  • User session recordings (PostHog, Hotjar, etc.)
  • Manually recorded user workflows
  • Customer support session recordings
  • User onboarding flows

​
What AI Learns from Recordings

  • User Interaction Patterns: How users navigate and interact with your app
  • Common Workflows: Typical user journeys and processes
  • UI Element Usage: How different interface elements are used
  • Error Handling: How users encounter and resolve issues

​
Manual Knowledge

Add custom documentation and context that isn’t available in other sources.

​
Creating Manual Knowledge

Document specific aspects of your application:
  • Business Logic: Custom rules and processes
  • Integration Details: Third-party service integrations
  • Edge Cases: Unusual scenarios and their handling
  • Configuration Details: Environment-specific settings

​
Best Practices for Manual Knowledge

  • Be Specific: Include detailed information about complex workflows
  • Use Examples: Provide concrete examples of expected behavior
  • Update Regularly: Keep documentation current with application changes
  • Focus on Testing: Emphasize information relevant to automated testing

​
Integration with AI Testing

​
How AI Uses Knowledge

The AI testing engine leverages your knowledge base to:
  • Understand Context: Comprehend the purpose and function of UI elements
  • Generate Realistic Tests: Create tests that follow actual user patterns and logical user journeys
  • Make Intelligent Decisions: Choose appropriate actions and better element selectors based on application knowledge
  • Provide Better Error Messages: Give meaningful feedback when tests fail and help avoid common testing pitfalls
  • Create Contextual Assertions: Generate more meaningful test validations based on application context