Skip to main content
NagaAI supports audio input for Gemini models via the input_audio content block in Chat Completions. For other models, use text/image/file in Chat Completions and the dedicated Speech/Transcription/Translation endpoints for audio processing.

​
How to Send Audio

​
Using Base64-Encoded Audio

Encode your audio file as base64 and provide it in the input_audio.data field: 
{
  "model": "gemini-2.5-flash",
  "messages": [
    {
      "role": "user",
      "content": [
        { "type": "text", "text": "Tell me what is said in this audio." },
        {
          "type": "input_audio",
          "input_audio": { "data": "BASE64_AUDIO", "format": "wav" }
        }
      ]
    }
  ]
}

​
Supported Audio Types

  • audio/wav
  • audio/mp3

​
Discover Supported Models

You can see which models accept audio input on the NagaAI Models page (audio filter).