Hi everyone, I’ve set up a VAPI assistant through the dashboard, but I’m running into issues with voice transcription. The audio-to-text conversion isn’t reliable—I've tested several models, but none consistently work well. My primary language is German. When I check the logs, it seems like the voice recording doesn’t actually capture my voice. I’ve tried different browsers, but the issue persists. Does anyone have suggestions for improving voice detection or troubleshooting this? Thanks in advance!

@HugoTavuk Hello! I'm sorry to hear you're experiencing issues with voice transcription. Here are a few suggestions that might help improve voice detection and troubleshoot the problem: 1. **Microphone Check**: Ensure your microphone is properly configured and working. Test it with other applications to confirm it's not a hardware issue. 2. **Environment**: Try to use the VAPI assistant in a quiet environment to minimize background noise that could interfere with voice detection. 3. **Pronunciation**: Since you're using German, make sure that you're speaking clearly and at a moderate pace to help the system better recognize your speech. 4. **Browser Permissions**: Double-check that your browser has the necessary permissions to access your microphone. 5. **Audio Settings**: Adjust the microphone sensitivity settings in your system or browser to ensure it's picking up your voice correctly. 6. **Update Browsers**: Ensure that you are using the latest version of your browser, as updates often include improvements to audio processing. 7. **Alternative Devices**: If possible, try using a different device to see if the issue is device-specific. 8. **Contact Support**: If none of these steps resolve the issue, it might be a more technical problem that requires support from the VAPI team. If you've already tried these steps without success, please let me know, and we can explore further options or escalate this to the support team for more in-depth assistance.

I have the same problem! I tried more devices, but it is not just about the device as the transcript of AI is also not precise. I am using deepgram transcriber , nova-2 model and my language is Czech. The voice is also Czech. Is there some possibility of using API connection to deepgram with more precise transcription to Czech language, please? What do you recommend me? If you need more info from my side, do not hesitate to ask me what do you need. Thank you!

Hey @HugoTavuk To help track down this issue, could you share: - The call ID - When exactly this happened (the timestamp) - What response you expected to get - What response you actually got instead This would really help me figure out what went wrong!

Seems like this was a browser issue that is persistent, but can be solved by restarting my laptop. So it's fine for now. Thanks!