How to Read Test Logs Step by Step
Test logs show every action the AI took during exploration, in chronological order. Each step shows what happened, when it happened, what the page looked like, and what evidence was collected. Reading logs helps you understand what the AI did and verify findings.
Log Structure
Each step in the log contains:
- Step number - Sequential action number
- Action type - What the AI did (click, type, navigate, etc.)
- Target - What element was interacted with
- Timestamp - When the action occurred
- Screenshot - What the page looked like
- Evidence - Console logs, network errors, etc.
- Status - Success, failed, skipped, or blocked
Step Types
Navigate
The AI loaded a new page:
Step 1: Navigate
Target: https://app.com
Status: ✓ Success
Duration: 2.3s
Click
The AI clicked an element:
Step 3: Click
Target: "Sign Up" button
Selector: button[data-testid="signup-btn"]
Status: ✓ Success
Duration: 0.5s
Type
The AI filled in a form field:
Step 5: Type
Target: Email input field
Value: test@example.com
Status: ✓ Success
Duration: 1.2s
Wait
The AI waited for something:
Step 7: Wait
Reason: Waiting for page to load
Duration: 1.8s
Screenshot
The AI captured the current state:
Step 9: Screenshot
Reason: After form submission
Captured: Current page state
Understanding Step Details
Screenshots
Each step includes a screenshot showing:
- What the page looked like at that moment
- What element was interacted with (highlighted)
- Visual state before and after actions
Screenshots help you:
- See what the AI saw
- Understand context around actions
- Verify issues are real
- Debug problems
Evidence
Steps may include evidence:
- Console logs - JavaScript errors or warnings
- Network logs - Failed requests, response codes
- DOM snapshots - HTML structure at that moment
- Error messages - Any errors that occurred
Timestamps
Timestamps show:
- When each action occurred
- How long each action took
- Total exploration duration
- Time between actions
Reading Logs Effectively
Start with Summary
Before diving into logs:
- Check the summary - pages explored, issues found
- Look at the issues list - what problems were detected
- Then read logs to understand context
Focus on Issue Steps
When reviewing issues:
- Jump to the step where issue was detected
- Look at steps before and after
- Review screenshot and evidence
- Understand the sequence that led to the issue
Use Search/Filter
Most log viewers allow:
- Searching for specific actions
- Filtering by status (failed, blocked, etc.)
- Jumping to specific step numbers
- Expanding/collapsing sections
Common Log Patterns
Successful Flow
Step 1: Navigate → Homepage
Step 2: Click → "Sign Up" button
Step 3: Type → Email field
Step 4: Type → Password field
Step 5: Click → "Submit" button
Step 6: Navigate → Dashboard (redirected)
✓ All steps successful
Failed Flow
Step 1: Navigate → Homepage
Step 2: Click → "Sign Up" button
Step 3: Type → Email field
Step 4: Type → Password field
Step 5: Click → "Submit" button ❌ FAILED
Error: Console error detected
Evidence: "Cannot read property 'value' of null"
Blocked Flow
Step 1: Navigate → Homepage
Step 2: Click → "Sign Up" button
Step 3: Type → Email field
Step 4: Type → Password field
Step 5: Click → "Submit" button ⏸ BLOCKED
Reason: CAPTCHA detected
Action: Cannot proceed automatically
What to Look For
Issues
When reviewing logs for issues:
- Failed steps - Actions that didn't complete
- Console errors - JavaScript errors logged
- Network errors - Failed requests
- Unexpected behavior - Steps that didn't do what was expected
Flow Understanding
To understand what the AI did:
- Follow the sequence of actions
- See how it navigated through your app
- Understand what paths it explored
- Identify what it might have missed
Performance
To check performance:
- Look at step durations
- Identify slow actions
- Find long wait times
- See overall exploration time
Tips for Reading Logs
- Start from issues - Jump to steps where issues were found, then read context
- Use screenshots - Visual context helps understand what happened
- Check timestamps - Long pauses might indicate problems
- Look for patterns - Repeated failures suggest systematic issues
- Compare steps - See before/after states to understand changes