VAEEncodeAudio - ComfyUI Built-in Node Documentation

Complete documentation for the VAEEncodeAudio node in ComfyUI. Learn its inputs, outputs, parameters and usage.

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

The VAEEncodeAudio node converts audio data into a latent representation using a Variational Autoencoder (VAE). It takes audio input and processes it through the VAE to generate compressed latent samples that can be used for further audio generation or manipulation tasks. The node automatically resamples audio to match the VAE’s expected sample rate if needed before encoding.

Inputs

Parameter	Data Type	Required	Range	Description
audio	AUDIO	Yes	-	The audio data to encode, containing waveform and sample rate information
vae	VAE	Yes	-	The Variational Autoencoder model used to encode the audio into latent space

Note: The audio input is automatically resampled to match the VAE’s expected sample rate (default: 44100 Hz) if the original sample rate differs from this value. If the input audio is None (e.g., the source video has no audio track), the node will raise an error.

Outputs

Output Name	Data Type	Description
LATENT	LATENT	The encoded audio representation in latent space, containing compressed samples

---

Source fingerprint (SHA-256): 5c65a853a121a557a5342d337c27e2ea3acff48e6b911bdaf27c89d4c7dc01b6