#Error while configuring nova 3

1 messages · Page 1 of 1 (latest)

novel rose
#

I see the following error when I configure nova 3 as the TTS model:
{"type":"Error","description":"Conflicting values for agent.language and agent.listen.provider language. Please remove the deprecated agent.language field.","code":"INVALID_SETTINGS"}

Here is the setting:
"listen": {
"provider": {
"type": "deepgram",
"model": "nova-3",
"language": "multi"
}
}

Please note that my goal is to use hindi and english as the language.

tawny sandalBOT
#

Hi @novel rose. I'm kapa.ai, your friendly support bot. I will try to answer your question by searching through the knowledge base. :wave:

#

The error is telling you that agent.language (top-level agent setting) conflicts with the language you’re passing under agent.listen.provider. The fix is to remove agent.language from your Settings payload and rely on the STT-side language configuration instead.

From the Voice Agent settings spec:

Your error message explicitly says:

Please remove the deprecated agent.language field.

So your Settings should look like this (key parts only):

{
  "type": "Settings",
  "audio": {
    "input": { "encoding": "linear16", "sample_rate": 24000 },
    "output": { "encoding": "linear16", "sample_rate": 24000 }
  },
  "agent": {
    // REMOVE agent.language entirely
    "listen": {
      "provider": {
        "type": "deepgram",
        "model": "nova-3",
        "language": "multi"
      }
    },
    "think": { /* your LLM config */ },
    "speak": {
      "provider": {
        "type": "deepgram",
        "model": "aura-2-thalia-en"  // or another TTS voice you prefer
      }
    }
  }
}

This configuration:

#
  • Avoids the conflict by omitting agent.language.
  • Uses model: "nova-3" with language: "multi" on the listen provider, which is the documented way to enable multilingual code-switching (e.g., English + Hindi) with Nova‑3. [Multilingual codeswitching; Nova-3]

If you still see the error after removing agent.language, the knowledge sources don’t provide further troubleshooting steps specific to that message, so I can’t say more with confidence.

Note: you can ask me a follow up question by @ mentioning me again :speech_balloon: