----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Orchestration Models]() * [Glossary]() * [Introduction]() * [Server URLs]() -----------------------------

### Answer It seems that Vapi's backchanneling feature is active while your assistant is running a function tool. This is happening because backchanneling is designed to make the conversation sound more natural by providing verbal feedback like "Yeah" and "Aha" during pauses. If you want to disable this behavior, you can adjust the

backchannelingEnabled

setting to

false

when creating or updating your assistant. Here's an example of how to disable backchanneling:

json
{
    "name": "Your Assistant Name",
    "backchannelingEnabled": false,
    "firstMessage": "Hello! How can I assist you today?",
    "serverUrl": "https://your-server-url.com",
    "model": {
        "provider": "openai",
        "model": "gpt-4-turbo"
    },
    "voice": {
        "provider": "azure",
        "voiceId": "andrew"
    }
}

[(1)]() , [(2)]() By setting

backchannelingEnabled

to

false

, the assistant will no longer interject with backchannel cues during function tool operations.