https://vapi.ai/ logo
Join Discord
Powered by
AI can't hear short answers like yes, ok etc
# support
j
AI often struggles to recognize short words such as "yes" or "ok." It is configured to talk in Italian (maybe is part of the problem). call id: 6375b822-08c5-45a6-889c-9fc8a647aa16
v
@JollyMask I understand that recognizing short words like "yes" or "ok" can sometimes be challenging for voice recognition systems, especially when configured for a specific language like Italian. To assist you better, could you please provide more details on the issue? Are you experiencing this problem during live interactions, or is it occurring with a specific set of voice data? Any additional context would be helpful to address your concern effectively.
You can continue this conversation by mentioning me in the message.
Got it. Tagging @User for additional help.
j
i tried changing the startspeaking plan this way as suggested in another post: "startSpeakingPlan": { "transcriptionEndpointingPlan": { "onPunctuationSeconds": 0.4 }, "waitSeconds": 0.3, "smartEndpointingEnabled": true }, but still doesn't catch short responses most of the time
any help would be much appreciated 🙏
@Shubham Bajaj
s
@JollyMask smartEndpointingEnabled only works for English transcriptions, you need to set it to false. Regarding your startSpeakingPlan try first with the waitSeconds to 0.8 and then add the transcriptionEndpointingPlan if required and also set the Deepgram's endpointing to 300ms.
Copy code
json
    "startSpeakingPlan": {
      "waitSeconds": 0.8,
      "smartEndpointingEnabled": false,
      
    }
Copy code
json
  "transcriber": {
      "model": "nova-2-general",
      "keywords": ["si", "pronto", "ok", "no"],
      "language": "it",
      "provider": "deepgram"
      "endpointing": 300
  }
Do let me know how it goes.
j
I'll try
Ok, I made three attempts, setting smartEndpointingEnabled to false and adjusting the parameters as indicated, but it still struggles to recognize short words 1st call id: c63cccdc-7228-4ba5-96e1-b72c11fab692 settings:
Copy code
"startSpeakingPlan": {
      "waitSeconds": 0.8,
      "smartEndpointingEnabled": false,
      
    }
2nd call id: dc1ec371-51c8-4bf1-b804-bc0fa82ac6a4 settings:
Copy code
"startSpeakingPlan": {
      "transcriptionEndpointingPlan": {
        "onPunctuationSeconds": 0.4
      },
      "waitSeconds": 0.8,
      "smartEndpointingEnabled": false
    }
3rd call (Too slow with the responses and unresponsive, it’s hard to have a conversation with it) id: 5084e308-4b04-486e-b145-88bbc1aefe01 settings:
Copy code
"transcriber": {
      "model": "nova-2-general",
      "language": "it",
      "provider": "deepgram",
      "keywords": [
        "si",
        "pronto",
        "ok",
        "no"
      ],
      "endpointing": 300
    },
    "startSpeakingPlan": {
      "transcriptionEndpointingPlan": {
        "onPunctuationSeconds": 0.4
      },
      "waitSeconds": 0.8,
      "smartEndpointingEnabled": false
    }
any suggestions?
s
@JollyMask Try this endpointing plan, and can you share your expectations and thought process for setting it to 0.4 seconds so i can get to understand what exactly your looking for?
Copy code
json
    "startSpeakingPlan": {
      "waitSeconds": 0.8,
      "smartEndpointingEnabled": false,
      
    }
  "transcriber": {
      "model": "nova-2-phonecall",
      "language": "it",
      "provider": "deepgram"
      "endpointing": 300
  }
j
message has been deleted
I set it to 0,4 because I see this post where this guy was having the same problem and you suggested this
if i set the model to nova-2-phonecall i receive this error 
{"message":["assistant.transcriber.language must be one of the following values for nova-2-phonecall model: en, en-US"],"error":"Bad Request","statusCode":400}
I think it doesn't support italian
My goal is to make the AI able of responding to short answers, words like "si (yes), ok, no ........"
?
s
@JollyMask You can use Nova2 in general, it's not an issue. Did you get a chance to try out the recent config? If so and it's not working, could you share the call ID so I could take a look?
m
@JollyMask Did it work for you?
Im having issues especially with si not being detected
s
@mohamed please create a new #1211483291191083018 ticket and tag me over there with the following info: - 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!
j
still having some issues
s
Hey @JollyMask can you please create a #1211483291191083018 ticket and list down your issue over there and ping me as well !!
i
ciao, sei italiana/o? se si, sentiamoci!!! necessito di un supporto con il nostro agente. se offri servizi o consulenza possiamo collaborare.
j
ti ho scritto
s
@intentionandflow what help do you require? Can you ping me to your #1211483291191083018 ticket.
d
Hey @Shubham Bajaj @JollyMask , we faced the same issue with short answers in Italian. Did you find any decent Assistant configuration to lower the chances of it happening?
s
Hey @Dmytro can you please create a new #1211483291191083018 ticket for your issue, and To help track down this issue, could you share the following in #1211483291191083018 ticket: - 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 us figure out what went wrong!
d
Hey @Shubham Bajaj, for now we achieved the result that we are satisfied with. If we face any issue in the future, we will create a separate ticket. Thanks for your answer
m
@Dmytro can you share, how you achieved the goal? i have same issue here and i'm looking for solution 🙂
k
This ticket has been marked as solved, and as a result, you will no longer receive any further responses. Kindly create a new support ticket to escalate your inquiry. 
5 Views
PreviousNext
Powered by