Troubleshooting

This guide covers common issues you may encounter when using the Chat App and how to resolve them.

Common Issues

Issue: No Response from AI

Symptoms:

• Message sends but no response appears
• Streaming indicator spins indefinitely
• Connection seems to hang

Cause: Usually caused by API key issues, network problems, or provider service outages.

Solution:

1. Verify your API key is correctly configured for the selected provider
2. Check your network connection
3. Verify the provider's service status:
   • Anthropic Status
   • OpenAI Status
4. For LM Studio, ensure the application is running and the model is loaded
5. Try switching to a different provider or model

Issue: Streaming Stops Mid-Response

Symptoms:

• Response starts streaming but stops unexpectedly
• Partial message appears then freezes
• No error message displayed

Cause: May be caused by token limits, network interruption, or server-side errors.

Solution:

1. Wait a few moments - the stream may resume
2. If using max tokens setting, try increasing the limit
3. Refresh the page and the partial message should be saved
4. Try sending a shorter prompt to avoid token limits
5. Check if the provider is experiencing issues

Issue: LM Studio Not Connecting

Symptoms:

• LM Studio option shows but doesn't work
• "Connection refused" or timeout errors
• Model selection dropdown is empty for LM Studio

Cause: LM Studio server may not be running or is on a different port.

Solution:

1. Open LM Studio application on your computer
2. Load a model in LM Studio
3. Start the local server (usually runs on http://localhost:1234)
4. Verify the server is running by visiting the URL in your browser
5. Check that no firewall is blocking the connection

Issue: Conversations Not Saving

Symptoms:

• Conversations disappear after page refresh
• Messages are lost
• New conversation button creates duplicates

Cause: Database connection issues or sync problems.

Solution:

1. Check your browser's console for RxDB errors
2. Verify you're logged into the correct organization/workspace
3. Try clearing browser cache and local storage (last resort - may lose local data)
4. Ensure the RxDB server is running if replication is enabled
5. Check if you have sufficient storage space in the browser

Issue: Tool Calls Failing

Symptoms:

• Tool call indicators show error state (red border)
• AI mentions using a tool but results don't appear
• Tool execution hangs in "pending" state

Cause: MCP Gateway connection issues or tool configuration problems.

Solution:

1. Verify MCP Gateway is running and accessible
2. Check that the required tools are registered with the gateway
3. Look at the expanded tool call details for specific error messages
4. Ensure you have permissions to execute the requested tools
5. Try refreshing tools by restarting the conversation

Issue: Markdown Not Rendering Correctly

Symptoms:

• Code blocks appear as plain text
• Markdown syntax is visible instead of formatted
• Syntax highlighting not working

Cause: May be a browser compatibility issue or rendering library problem.

Solution:

1. Ensure you're using a supported browser (Chrome, Firefox, Safari, Edge)
2. Try refreshing the page
3. Check if the issue persists in a different browser
4. Report the issue with the specific markdown content that fails

Issue: Model Selector Not Showing All Models

Symptoms:

• Expected models are missing from dropdown
• Only one provider appears
• Dropdown is empty

Cause: Missing API keys or configuration for certain providers.

Solution:

1. Verify API keys are configured for all desired providers
2. Check that keys have the correct permissions
3. For enterprise accounts, ensure your plan includes the models you need
4. LM Studio models only appear when the server is running

Performance Issues

Issue: Slow Response Times

Symptoms:

• Long delays before streaming starts
• Responses take much longer than expected
• Interface feels sluggish

Cause: Network latency, model choice, or prompt complexity.

Solution:

1. Try a faster model (e.g., Claude Haiku instead of Claude Opus)
2. Check your internet connection speed
3. Reduce prompt length if possible
4. For LM Studio, ensure your hardware can handle the loaded model
5. Consider using a wired connection instead of WiFi

Getting Help

If you can't resolve your issue:

1. Check the FAQ for additional answers
2. Review the Features documentation for proper usage
3. Contact support with:
   • Description of the issue
   • Steps to reproduce
   • Screenshots if applicable
   • Browser and device information
   • Selected provider and model
   • Any error messages from the browser console