# Voice Interaction

OMate supports voice output functionality and voice chat, making AI character responses more lively and natural.

Video tutorial

# Voice Playback

1. When ① voice model is configured ② and voice parameters are configured in the character settings, a voice icon will appear in the chat interface. Click on it to play the voice.

# Voice Settings

Configure voice parameters in the settings page:

1. Select a voice model
2. Set voice parameters: please configure according to the voice model's documentation
3. Go to the character list, find the character card that needs voice settings, swipe left to edit
4. Enter the relevant parameters in the voice parameters field on the edit page, save
5. Enter the character conversation interface, click the voice icon in the top left of the text to play the voice

The specific parameters depend on the service provider's text-to-speech documentation. Taking Silicon Flow as an example:

1. Enter FunAudioLLM/CosyVoice2-0.5B in the voice model field in the API settings
2. Enter voice=FunAudioLLM/CosyVoice2-0.5B:alex in the voice parameters field of the character card, where alex is a preset voice. You can refer to the documentation to adjust:

# Voice Call

Once you have configured voice playback, if you also configure the speech recognition model, you will be able to use the voice call feature.

You can use the speech recognition model from the same service provider as your chat model, just enter the name directly. For example:

• Silicon Flow: iic/SenseVoiceSmall
• OpenAI: whisper-1

You can also specify a separate service provider for speech recognition by clicking the ➕ in the settings item, where you can separately fill in the api address, key, and parameters.