Skip to main content
Common issues and how to resolve them.

​
Test Failures

Symptom: Test fails because the agent can’t locate a button, link, or elementCommon causes:
  • UI changed (element was moved, renamed, or removed)
  • Element loads dynamically and wasn’t ready
  • Element is hidden or off-screen
  • Modal or popup blocking the element
Solutions:
  1. Watch the video - See exactly what the agent saw
    • Was the element visible on screen?
    • Did the page finish loading?
    • Was anything blocking it?
  2. Make the AI step more specific - Better description helps
    • Instead of “Click submit”, try “Click the green submit button in the bottom right”
    • Include color, position, or surrounding context
    • Mention unique text or labels
  3. Add wait steps - Give elements time to appear
    • Add a wait step before the interaction
    • Wait for specific content to load
  4. Check for popups - Something might be in the way
    • Watch the video for modals or notifications
    • Add steps to close blocking elements first
Symptom: Check step assertion failedCommon causes:
  • Check is too strict or specific
  • Content is dynamic (dates, IDs, random data)
  • Timing issue (content not loaded yet)
  • Application behavior actually changed
Solutions:
  1. Watch the video - What did the page actually show?
    • Compare to what you expected
    • Is the app behaving correctly?
    • Did the check run too soon?
  2. Review the check - Is it realistic?
    • Is the check too specific?
    • Does it look for exact text that might change?
    • Can you make it more flexible?
  3. Add a wait step - Give content time to appear
    • Add wait before the check
    • Wait for loading indicators to disappear
  4. Update the check - If app behavior changed
    • Edit the check prompt to match current behavior
    • Make it less strict if needed
    • Remove if no longer relevant
Symptom: Test fails with timeout or takes too longCommon causes:
  • Application is slow to load
  • Network issues
  • Waiting for something that never appears
  • Infinite loading states
Solutions:
  1. Watch the video - What was happening when it timed out?
    • Was the page still loading?
    • Was there a spinner or loading indicator?
    • Did the agent keep trying the same action?
  2. Check network logs - Look for slow or failing requests
    • Are API calls taking too long?
    • Any failed requests (500, 404)?
    • Network connectivity issues?
  3. Add wait steps - Give slow content time to load
    • Add explicit waits before time-sensitive steps
    • Wait for loading indicators to disappear
  4. Test your app manually - Is it actually slow?
    • Does your app load quickly when you use it?
    • Are there known performance issues?
    • Check if the environment is down or degraded
Symptom: Test redirected to login or can’t access protected pagesCommon causes:
  • Browser profile expired
  • Session timeout
  • Profile not selected
  • Authentication changed
Solutions:
  1. Watch the video - Did it redirect to login?
    • Check if login page appeared during test
    • Look for “session expired” messages
  2. Update browser profile - Refresh session
    • Go to Profiles
    • Click “Resume” on profile
    • Re-authenticate completely
    • Save profile
  3. Verify profile is selected - Check test configuration
    • Open the test
    • Check Profile dropdown at top
    • Select correct profile
    • Re-run test
  4. Check network logs - Look for auth errors
    • See any 401 or 403 responses?
    • Failed authentication API calls?

​
Scheduling Issues

Symptom: Expected test run missing from historyCheck:
  • ✅ Schedule is active (not paused)
  • ✅ Timezone is correct for your location
  • ✅ Tests in schedule are enabled
  • ✅ No system maintenance during scheduled time
Solutions:
  • Verify next run time in schedule details
  • Check timezone setting carefully
  • Enable disabled tests
  • Review schedule history for errors
Symptom: Entire batch failed at onceCommon causes:
  • Application was down
  • Browser profiles expired
  • Environment changed (URLs, credentials)
  • Network connectivity issues
Solutions:
  • Check application was accessible at scheduled time
  • Update all browser profiles
  • Verify environment URLs in tests
  • Review test run error messages
  • Test manually to reproduce
Symptom: Displayed next run time doesn’t match expectationCauses:
  • Timezone confusion
  • Daylight saving time transition
  • Schedule was recently edited
Solutions:
  • Double-check timezone setting
  • Account for DST if applicable
  • Verify schedule configuration
  • Compare to your local time carefully

​
Browser Profile Issues

Symptom: Tests using profile still see login pageSolutions:
  1. Resume profile
  2. Re-authenticate completely
  3. Navigate through app to verify login
  4. Save profile
  5. Re-run test
Set calendar reminders to refresh profiles weekly or monthly
Symptom: Test runs as different user than expectedCheck:
  • Correct profile selected in test config
  • Profile hasn’t been overwritten
  • Profile was saved after correct login
  • No name confusion between profiles
Solutions:
  • Verify profile selection
  • Rename profiles for clarity
  • Create new profile if corrupted
  • Test profile in isolation first
Symptom: Save button errors or doesn’t workTry:
  • Ensure fully logged in before saving
  • Close and resume session
  • Check browser console for errors
  • Try different browser if possible
  • Contact support with error details

​
Swarm Issues

Causes:
  • Authentication failed
  • Starting URL inaccessible
  • Application requires specific setup
  • Guidelines too restrictive
Solutions:
  • Test credentials manually first
  • Verify URL is accessible
  • Use browser profile for complex auth
  • Add focus area description
  • Check Swarm run logs for errors
Symptom: Generated tests are trivial or incorrectImprove with:
  • More detailed focus area description
  • Clear guidelines (what to avoid)
  • Documentation URLs
  • Lower branching factor
  • Better authentication (use profile)
Solutions:
  • Lower target test count
  • Reduce branching factor
  • Add specific focus area
  • Increase max concurrency
  • Add guidelines to avoid problem areas

​
Video Upload Issues

Verify:
  • Video is under 10 minutes
  • Format is supported (MP4, MOV, WebM)
  • Video has clear visuals
  • File isn’t corrupted
Try:
  • Re-export video at lower quality
  • Convert to MP4 format
  • Trim video to shorter length
  • Record at lower resolution
Remember: AI extraction isn’t perfectSolutions:
  • Edit steps manually to fix them
  • Delete incorrect or unnecessary steps
  • Add missing steps
  • Rewrite step goals to be more specific
  • Consider building from scratch if too many issues

​
General Debugging Strategy

1

Watch the video

See exactly what the agent did and what went wrong
  • Where did it fail?
  • What was on the screen?
  • Did the app behave correctly?
2

Check the timeline

Review each action the agent took
  • What was it trying to do?
  • What did it see at each step?
  • Where did it get stuck?
3

Review logs

Check console and network tabs
  • Any JavaScript errors?
  • Failed API calls?
  • Authentication issues?
4

Test manually

Try the same workflow in your browser
  • Does the application work correctly?
  • Are there any errors?
  • Did the UI change recently?
5

Fix and verify

  • Update test steps based on findings
  • Run test multiple times to verify
  • Monitor for stability
  • Add to schedule once stable

​
Getting Help

Still stuck? We’re here to help.

Email Support

[email protected]

Slack

Message us in your workspace channel
Include your test run ID and what you’ve already tried so we can help you faster.