guys, my assistant shows 0.23$ per minute but when i talk for 1.5 minutes it deducts 60+ cents its draining my bank account the overview shows nothing about my assistant and does anyone know how we can get analysis report and summary using squads? and did anyone figure out how to log vapi calls to gh

@Shahzaib.aiii I'm sorry to hear that you're experiencing issues with billing and accessing information about your assistant. Let's address your concerns one by one: 1. **Billing Discrepancy**: If you're noticing a discrepancy in billing, it's important to reach out to our support team directly so they can look into your account details and resolve any issues. You can email them at [support@vapi.ai](support@vapi.ai) with your account information and a description of the problem. 2. **Assistant Overview**: If the overview is not showing information about your assistant, this could be a technical glitch or a problem with the account settings. Again, contacting support will be the best course of action to get this resolved. 3. **Analysis Report and Summary Using Squads**: To get an analysis report and summary using squads, you would typically need to use the specific features provided by VAPI for call analysis. However, since the context does not provide details on how to do this with squads, I recommend checking the Quickstart Guides, Examples, or FAQ sections for guidance, or reaching out to support for more detailed instructions. 4. **Logging VAPI Calls to GitHub**: Logging API calls to GitHub is not a standard feature provided by VAPI, and it would require custom implementation. You might need to set up a webhook or an event listener that captures the VAPI call data and then uses the GitHub API to log that information into a repository. For specific instructions, you may need to consult with a developer or reach out to support for more info on this. If you need immediate assistance with any of these issues, please don't hesitate to send an email to our support team. Would you like me to assist you with anything else at this moment?

not github i want ot log calls in gohighlevel using make

@Shahzaib.aiii Could you share recent call IDs so I can take a look? Also, regarding logging calls to GHL, you can use the end-of-call report and GHL webhooks to determine the structure and map the response to the required fields.

@Shubham Bajaj 33fcd146-eaa2-46f6-89c9-10abc2267f02

@Shahzaib.aiii The model temperature is currently set to 1, which enhances productivity and creativity but may lead to hallucinations. I recommend adjusting your model temperature to a range between 0 and 0.3; this should resolve your issue. Please give it a try for 2-3 calls and let me know the outcome.

@Shubham Bajaj 11cd711a-44ec-4279-9700-e757267b8fb6 check this, my client called in front of me, the agent was repeating the same lines, "do you want to book an existing appointment or a new one?" and the client kept saying new one, but in recording we cant hear the agent malfunctioning! is there anyway to reboot the vapi or smthg ihave a lot clients and after some using the agents starts malfunctioning! or is it eleven labs voices? can a particular voice cause a malfunction! i need to learn this in detail I have clients and its getting hard for me to keep up w the quality

@Shahzaib.aiii can you try one more call without the KB and then let me know how it goes? https://cdn.discordapp.com/attachments/1331035264751112202/1334078796617355284/Screenshot_2025-01-29_at_2.02.28_PM.png?ex=679b3933&is=6799e7b3&hm=7219d2e47a68992c6ce3385a3beeae091529126d33b3ad6bfff88f3b6014db35&

@User I did, its hallucinating, what do i do its dead end for me

@Shahzaib.aiii

logs

[1] 🔵 16:28:40:184 Say Request Saying 19... (say: Hi! is this Cherie?) { "role": "bot", "time": 1738340921408, "source": "", "endTime": 1738340922628, "message": "Hi. Is this Sherry?", "duration": 1220, "secondsFromStart": 1.68 }, - Try Changing your first message it will be resolved. Because your message is correct but voice provider is hallucinating here things are going a bit off. [2] 🔵 16:28:43:006 [user LOG] Transcriber output: Yes. 🔵 16:28:43:014 [user LOG] Endpointing timeout 1014ms (rule:

smart

) 🔵 16:28:44:032 AzureOpenAIRequest. Messages: [ { "role": "system", "content": "You are Abraham, a dedicated cold caller for *..." }, { "role": "assistant", "content": "Hi. Is this Sherry?" }, {"role": "user", "content": "Yes."} ] 🔵 16:28:47:094 [user LOG] Endpointing timeout 925ms (rule:

smart

) 🔵 16:28:48:023 OpenAIRequest Request. Messages: [ { "role": "system", "content": "You are Abraham, a dedicated cold caller for *..." }, { "role": "assistant", "content": "Hi. Is this Sherry?" }, {"role": "user", "content": "Yes. Yes. Yes."} ] 🔵 16:28:50:200 [user LOG] Transcriber output: And who's calling? 🔵 16:28:50:215 OpenAIRequest Request. Messages: [ { "role": "system", "content": "You are Abraham, a dedicated cold caller for *..." }, { "role": "assistant", "content": "Hi. Is this Sherry?" }, { "role": "user", "content": "Yes. Yes. Yes. And who's calling?" } ] 🔵 16:29:05:550 [user LOG] Transcriber output: Sure. - This was an pure endpointing issue where it took the model little longer to identify the user has stopped speaking, sometimes models takes little longer to respond.

Ill have to listen it again, but i am talking about the delay ,v high latency

This was because of end pointing model, and startSpeakingPlan used. Use the following config instead and then you wont' observer the percieved latency.

json
    "start_speaking_plan": {
      "waitSeconds": 0.8,
      "smartEndpointingEnabled": true
    }

json
    "transcriber": {
      "model": "nova-2",
      "language": "en",
      "provider": "deepgram",
      "endpointing": 300
    },

Okay thankyou, Also is their any issue w connection of twilio w canada or smthg,i have two clients in canada and when ever i call myself the agent still works okay but when they call from there its a total disaster, it keeps on repeating the same lines like 6-7 times

@Shahzaib.aiii No there is no issue can you share the call ID so I can take a look?

sure, will will send you in a while. But i sent you one before as well.

"AI: Hello. This is Nevan from Ottawa Physiotherapy.\nUser: Just looking to book an appointment.\n", Can you check the call Id again or any recent call ID you have will work. I want you to get unblocked asap.

Yeah so in this call, i was in the call, and the agent kept saying how i can help you and my client kept on saying book an appointment, 5-6 times the agent repeated that, but when i listened to the recordings, i only heard my client repeating his words

Let us uphold the decorum of the forum by continuing to communicate in English.

Okay 👌

@Shahzaib.aiii deepgram changes have to be done using API.

In the make http module?

@Shahzaib.aiii Your right it's PATCH.

d6eeb629-4189-43ab-8df8-9bb844f420de

@Shahzaib.aiii looking into it, and I suggest creating new #1211483291191083018 for new issues from next time for SLA management.

How to change it to Trieve

We have migrated from Canonical to Trieve. You will need to attach the Knowledge Base using the API. You can learn more about it here: https://docs.vapi.ai/knowledgebase

What is the file ID