FishAudioVoiceSelector - ComfyUI Built-in Node Documentation - ComfyUI

The Fish Audio Voice Selector node selects a voice from the Fish Audio library for text-to-speech generation. You can pick one of the built-in preset voices, or choose “custom” to enter any voice model ID from fish.audio.

Inputs

Common Inputs

Parameter	Description	Data Type	Required	Range
`voice`	Choose a voice, or ‘custom’ to enter any fish.audio voice model ID.	DYNAMIC_COMBO	Yes	”Energetic Male (en)" "Friendly Women (en)" "Sarah (en)" "Verity (en)" "Polo (en)" "Adrian (en)" "E-girl (en)" "Narrator (en)" "Warm Conversational Voice (en)" "Warm Storyteller (en)" "Dramatic Character Male (en)" "News Narrator (zh)" "Lively Female (zh)" "Gentle Female (zh)" "Energetic Female (ja)" "Calm Female (ja)" "Calm Male (ja)" "custom”

The preset voice options cover English (en), Chinese (zh), and Japanese (ja) voices and do not require any additional inputs.

Custom Inputs

These inputs appear when voice is set to “custom”.

Parameter	Description	Data Type	Required	Range
`voice_id`	Voice model ID from fish.audio, e.g. the ID in https://fish.audio/m/<id>/. Default: empty string.	STRING	Yes	Any valid Fish Audio voice model ID

Note: When voice is set to “custom”, voice_id must not be empty after whitespace is trimmed; otherwise the node raises a “Custom voice ID is empty.” error. If an unrecognized voice option is passed, the node raises an “Unknown voice” error.

Outputs

Output Name	Description	Data Type
`voice`	The selected Fish Audio voice model ID. For a preset voice, the corresponding voice ID from the Fish Audio library is returned; for “custom”, the entered `voice_id` value is returned.	FISHAUDIO_VOICE

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

Source fingerprint (SHA-256): 4f99a58aa7e6054f58fe84e61e4e1008b17828bd97d71ef0a4009c4de4052bbd