Follow-up on Squad Initial Agent Transcription Iss...
# support
h
@Vapi @Vapi Support Bot I’m following up on the transcription issue I reported earlier with the Squad Initial Agent in Vapi Squads. The problem is that when a user speaks, their audio is not being transcribed. In my previous post, Vapi confirmed that this was an error on your side and mentioned that it was being fixed. Could you please provide an update on the status of this fix? This issue is blocking smooth interaction in my squad setup, so any timeline or workaround guidance would be greatly appreciated.
v
Thank you for your follow-up. While I don't have a direct update on the fix for the Squad Initial Agent transcription issue, I can help troubleshoot or suggest workarounds. Could you confirm if you have already tried configuring a fallback transcriber in your Squad setup (e.g., adding a secondary provider like Deepgram or Azure)? This can help maintain transcription continuity if the primary provider fails, as described in the [Vapi documentation](https://docs.vapi.ai/customization/transcriber-fallback-plan). If you haven't set up a fallback, I can guide you through the process or help review your current configuration for potential issues. Sources: - [Transcriber fallback configuration](https://docs.vapi.ai/customization/transcriber-fallback-plan) - [Debugging voice agents](https://docs.vapi.ai/debugging)
a
while waiting for their fix, you can try a quick workaround by switching the transcription provider (e.g. Deepgram ↔︎ AssemblyAI) or reinitializing the agent with a fresh config this sometimes restores user speech capture. If you want, I can quickly review your squad setup and pinpoint the exact issue or apply a temporary fix so your flow keeps working. @hamza-007
c
could you please link the other post for more context? thank you
@Vapi @Vapi Support Bot .....
c
we have sent a reply to that thread. i will put it here for your convenience: since the audio quality is a bit low, i would suggest allowing the use of keypad entry for languages so that the transcriber doesnt have to do much of the heavy lifting of deciphering the audio. for example, "for francais, press 1.... etc."