PikaMusicNode - ComfyUI Built-in Node Documentation - ComfyUI

This node generates music from a text description and optional lyrics, or covers a song you supply, with Pika. Choose "text to music" to write a new song at a length you set, or "cover" to re-record a song and keep its length.

Inputs

Parameter	Description	Data Type	Required	Range
`mode`	What the node does: `"text to music"` writes a new song, `"cover"` re-records a supplied song. Each mode reveals its own inputs below. Default: `"text to music"`.	DYNAMIC_COMBO	Yes	`"text to music"` `"cover"`

Text to Music Inputs

These inputs appear when mode is set to "text to music".

Parameter	Description	Data Type	Required	Range
`prompt`	What the music should sound like. Must not be empty. Default: "" (empty string).	STRING	Yes	1 to 4000 characters
`lyrics`	Optional lyrics. Section tags such as `[Verse]` and `[Chorus]` are allowed. Default: "" (empty string).	STRING	Yes	0 to 8000 characters
`duration`	Length of the song in seconds. Default: 30.0.	FLOAT	Yes	10.0 to 360.0 (step: 0.1)
`rewrite_prompt`	Expand a short or vague prompt with a model before generating. Default: False.	BOOLEAN	Yes	true false
`seed`	Seed for sampling. Default: 42.	INT	Yes	0 to 2147483647

Cover Inputs

These inputs appear when mode is set to "cover".

Parameter	Description	Data Type	Required	Range
`source_audio`	The song to cover. The result has the same length, clamped to 5 to 360 seconds.	AUDIO	Yes	-
`prompt`	What the cover should sound like. Must not be empty. Default: "" (empty string).	STRING	Yes	1 to 4000 characters
`lyrics`	Optional lyrics. Section tags such as `[Verse]` and `[Chorus]` are allowed. Default: "" (empty string).	STRING	Yes	0 to 8000 characters
`rewrite_prompt`	Expand a short or vague prompt with a model before generating. Default: False.	BOOLEAN	Yes	true false
`seed`	Seed for sampling. Default: 42.	INT	Yes	0 to 2147483647

Constraints:

The prompt must be between 1 and 4000 characters, and the lyrics 8000 characters or fewer.
Cover mode ignores duration: the output matches the length of source_audio, clamped to 5 to 360 seconds.

Outputs

Output Name	Description	Data Type
`AUDIO`	The generated song, or the cover of the supplied track.	AUDIO

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): e3051b2d0cfff72cf7e359d7383f4e0d133df0338ff42a00f0649387f504ec28