----------------------------- **While I generate an answer, here are some sources that might be helpful**: * [Dashboard]() * [Web Calling]() * [Outbound Calling]() * [Create Assistant]() -----------------------------

### Answer The JSON configuration you've shown is typically used for setting up various parameters of an assistant, such as the transcriber, voice, and language mappings. However, this JSON shouldn't be directly pasted into the system prompt field. In Vapi, the system prompt is designed to dictate the assistant's initial interaction and define its behavior during a session. This includes setting the context, instructions, and personality of the assistant. If you're looking to configure the assistant's transcriber and voice properties, you should do this through the assistant's setup options in the Vapi dashboard or via the API, not through the system prompt. For setting up or modifying an assistant's configuration on the dashboard: 1. Log into the Vapi dashboard. 2. Navigate to the 'Assistants' section. 3. Create a new assistant or edit an existing one. 4. Configure the transcriber, voice, and other settings by providing the appropriate JSON in the configuration sections. Alternatively, if you prefer using API calls, you can refer to relevant endpoints in the Vapi documentation for creating or updating an assistant, where you can pass your JSON configuration [(1)]() . If you need to set or update these configurations programmatically, here’s a sample API call using the POST method to create or update an assistant:

javascript
fetch('https://api.vapi.ai/assistant', {
  method: 'POST',
  headers: {
    'Authorization': 'Bearer YOUR_ACCESS_TOKEN',
    'Content-Type': 'application/json'
  },
  body: JSON.stringify({
    "transcriber": {
      "provider": "deepgram",
      "model": "nova",
      "languages": ["en", "de", "pl"]
    },
    "voice": {
      "provider": "11labs",
      "voiceId": "t0jbNlBVZ17f02VDIeMI"
    },
    "languageMappings": {
      "de": {
        "voiceId": "vj2Y0u1NaT1ayH6VYkYu"
      },
      "pl": {
        "voiceId": "7u8qsX4HQsSHJ0f8xsQZ"
      }
    }
  })
})
.then(response => response.json())
.then(data => console.log(data));

This script assumes you have proper authentication and the necessary permissions to perform such operations. Adjust the endpoint URL and method according to whether you're creating a new assistant or updating an existing one.

You can continue this conversation by mentioning me in the message.

@Rafal Kwiatkowski Sweet. Have a great day!