Audio nodes

Audio nodes cover voice and sound: text to speech, music, avatars, voice cloning and design, and audio tools that work with video.

Speech generation (TTS)

Open “Audio” in the “Add node” menu to create voice and music nodes (entries with no available model right now are hidden); avatars have their own entry in the same menu. In a speech generation node:

• Enter the text to be read aloud, choose a voice and settings, and generate a clip of speech;
• Sliders for speed, pitch and volume, and the “Pause” and “Tone” markup buttons, appear only when the selected model supports them;
• The maximum text length depends on the selected model.

Voice cloning and design

• The voice picker has four tabs: “Voice library”, “My voices”, “Favorites” and “Elements”;
• Clone voice: record yourself reading a sample text in the browser (about 10–20 seconds) to register a personal voice tied to the selected model, which you can then use for speech;
• Design voice: create a new voice from a description, with no reference audio needed;
• Whether there's a charge depends on the model: some charge a registration fee when the voice is created, some charge on the first synthesis, and some are free. Registration finishes in the background, and you can pick the voice once its status changes from “Registering” to “Ready”. Voices left unused for a long time are cleaned up automatically. A voice design preview may also be charged, so check the selected model's quote before you submit.

Audio tools

Many audio features actually live on the video node toolbar (see the “Video nodes” chapter): split audio & video, mix audio, vocal separation and more let you pull the sound out on its own or replace it, and most of them are free.

The “+” beside an audio node also has voice changer, voice isolator, generate from reference audio, and dubbing; choose one to create that node, connected automatically.