Unicode support
# support
m
In first message or prompt unicode characters are not supported. Whaat is the fix for that?
v
@mathews I understand that you're experiencing issues with Unicode characters in the first message or prompt. To address this, we can enable Unicode support in the input preprocessing settings. However, as a Voice API support bot, I don't have direct access to modify these settings. You would typically need to check the documentation of your Voice API provider, in this case, Azure, to understand how to configure Unicode support for voice inputs. If you have access to the configuration settings, look for an option that allows you to enable Unicode character support. If you're unable to find this setting or need further assistance, please reach out to Azure support for more detailed guidance on how to resolve this issue.
You can continue this conversation by mentioning me in the message.
m
we are sending hindi characters in the message but it is not picking up
s
can you please share call id.
m
1a90020c-051e-4efc-afbd-b6e435055960
also , can you please tell how to send the first_message from the custom llm. Because i couln't find that...
s
i couldn't find unicode characters can you share the timestamp.
this i haven't tested yet but can you try with firstMessageMode set to
"assistant-speaks-first-with-model-generated-message"
m
12:50:08:840. like we tried sending hello बादल, but it came out to be hello Baadal
s
this is not from Vapi, it was most probably done by SST as same is done for next sentence. Voice Input:
Copy code
🔵 07:20:08:839 Voice Input Formatted: "Hi, Badal. This is करन from UniversityLiving.", Original: "Hi, Badal. This is करन from UniversityLiving."
Transcript (as your lang is en-hi):
Copy code
🔵 07:20:12:996 `assistant` Final Transcript : Is Gurun from University Living.: 0.9135742