English

日本語
中文

English

日本語
中文

Appearance

Audio Mode Settings ​

Overview ​

In AITuberKit, you can use Audio Mode, which utilizes OpenAI's Audio API feature to respond with natural voice to text or voice input. This mode is provided as a different feature from the Realtime API mode.

Environment Variables:

bash
# Enable Audio Mode
NEXT_PUBLIC_AUDIO_MODE=false

# Use as the initial value for browser settings
NEXT_PUBLIC_OPENAI_API_KEY=sk-...

# Keep the key server-side instead
OPENAI_API_KEY=sk-...

# Audio mode input type (input_text or input_audio)
NEXT_PUBLIC_AUDIO_MODE_INPUT_TYPE=input_text

# Audio mode voice (alloy, coral, echo, verse, ballad, ash, shimmer, sage)
NEXT_PUBLIC_AUDIO_MODE_VOICE=alloy

AITuberKit proxies OpenAI requests through /api/ai/audio. When the browser-side API key is empty, the route uses the server-side OPENAI_KEY or OPENAI_API_KEY. To use a server-side key in a public deployment, set AITUBERKIT_SERVER_SECRET_ACCESS_MODE to protected or demo.

Supported Models ​

Audio Mode supports the following models:

  • gpt-audio-1.5
  • gpt-audio
  • gpt-audio-2025-08-28
  • gpt-audio-mini (default)
  • gpt-audio-mini-2025-12-15
  • gpt-audio-mini-2025-10-06

If you have a legacy gpt-4o-*-audio-preview model saved, it is automatically migrated to the corresponding gpt-audio or gpt-audio-mini model at startup.

Setup Method ​

To use Audio Mode, follow these steps:

  1. Select OpenAI as the AI service
  2. Set up the OpenAI API key
  3. Turn ON the Audio Mode
  4. Select the input type and voice as needed

Transmission Type Settings ​

In Audio Mode, you can choose from two transmission methods:

  • Text: Transcribes voice input with Web Speech API before sending
  • Voice: Sends voice data directly from the microphone to the Realtime API

Voice Type Settings ​

In Audio Mode, the following voice types are available:

  • alloy, coral, echo, verse, ballad, ash, shimmer, sage

Each voice has different characteristics, allowing you to select the optimal voice for your character.

Limitations ​

  • Currently only supports OpenAI's service
  • Cannot be used with External Linkage mode or Realtime API mode
  • May incur higher API usage fees than other modes