Hi there! đź‘‹ I understand you're having issues with the transcriber defaulting to English and not saving changes when switched to another language.
I don't have enough information about your specific setup to determine exactly why the changes aren't being saved. However, I can help with several potential causes and solutions for the transcriber defaulting to English.
## Potential Causes
There are several reasons why your transcriber might be defaulting to English:
1. **Model Limitations**: If you're using the Nova-3 model, it only supports English ('en' and 'en-US') according to [documentation](
https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1341463939246198784). For other languages, you'll need to use Nova-2 or Nova-2-general.
2. **Incorrect Configuration**: The language settings might not be correctly configured for your desired language. Double-check your language and accent settings in the Vapi configuration.
3. **Code Switching Issues**: If you're using Deepgram, its code switching feature only works for English and Spanish according to [their documentation](
https://discord.com/channels/1211482211119796234/1211483291191083018/threads/1308120552140964034).
## Troubleshooting Steps
Here are some steps you can take to resolve the issue:
1. **Verify Language Settings**: Ensure that the language settings are correctly configured for the languages you want to transcribe.
2. **Switch Transcribers**: Try different transcribers. Some users have resolved similar issues by switching from Deepgram to Azure or Talkscriber.
3. **Use the Right Model**: If using Nova-3, switch to Nova-2 or Nova-2-general for non-English languages.
4. **Check Browser Issues**: Clear your browser's cache and cookies, disable extensions, or try a different browser.
5. **Update SDK**: Consider updating the SDK to the latest version if you're using an older version.