Skip to main content

Troubleshooting

Common issues and solutions for The One AI Platform.

Jarvis Not Responding​

Symptoms: Jarvis chat shows a loading spinner or returns an error.

Solutions:

  1. Check your AI Platform subscription is active in Hub Billing
  2. Verify your usage quota hasn't been exceeded for the current period
  3. The upstream AI provider may be experiencing an outage — check the status page
  4. Try refreshing the page and starting a new conversation

Custom Agent Not Working as Expected​

Symptoms: A custom agent returns irrelevant or incorrect responses.

Solutions:

  1. Review the agent's system prompt — ensure it has clear instructions and context
  2. Check the knowledge base attached to the agent — missing or outdated documents affect quality
  3. Try a different model — some models perform better for specific tasks
  4. Test with simpler prompts first to isolate the issue

Usage Quota Exceeded​

Symptoms: AI features return "quota exceeded" errors.

Solutions:

  1. Check current usage in Settings → Usage Quotas
  2. Wait for the quota period to reset (monthly by default)
  3. Increase the quota limit in Settings if needed
  4. Review usage analytics to identify which team or agent is consuming the most

Prompt Template Not Producing Expected Output​

Symptoms: A prompt template generates inconsistent or incorrect results.

Solutions:

  1. Review variable placeholders — ensure all {{variables}} have values at runtime
  2. Test the template with the Preview function before deploying
  3. Add more specific instructions to the template's system context
  4. Check the model selection — ensure the template uses an appropriate model for the task

Studio App Build Errors​

Symptoms: A Studio app won't publish or shows errors in the builder.

Solutions:

  1. Check for disconnected blocks — all blocks must be connected in the flow
  2. Verify required configuration on each block (API keys, model selection, field mappings)
  3. Test the app in preview mode to see specific error messages
  4. Check that integration blocks have valid credentials

AI Gateway Rate Limiting​

Symptoms: Requests are being throttled or returning 429 errors.

Solutions:

  1. Check your tier's rate limits in Settings → Usage Quotas
  2. Spread requests over time if running batch operations
  3. Use the queue system for high-volume operations instead of direct API calls
  4. Contact support if your usage patterns require higher rate limits
ℹ️If you encounter an issue not listed here, contact support with the conversation ID (for Jarvis), agent ID (for custom agents), or app ID (for Studio apps).

Next Steps​