Build

Turn text into speech with Duaer Speech

In Duaer, Duaer Speech reads a text aloud in a preset voice or a voice you cloned, and returns the audio with its duration.

Read a text aloud

Search for Duaer Speech and connect it after the previous node.
In Text, enter what to read. You can use an expression to pick up copy written by an earlier step.
Open the Model dropdown and pick a speech model.
Open the Voice dropdown and pick a voice. Model Default uses the model's own voice; the list shows the Duaer preset voices first, then the voices you have cloned.
Select Execute step. The binary field data holds the audio, plus duration in seconds, model, voice, and text.

Use duration to time the picture, for example as data-duration in Duaer Video Render so the video runs as long as the narration.

Duaer Speech options

Speed: from 0.5 to 2. The default is 1.
Format: MP3 or WAV. The default is MP3.
Output Binary Field: the binary field that holds the audio. The default is data.

Use your own voice

Under Settings › My voices, select Clone a voice and read the script on the page for 10 to 30 seconds. When the status changes from Training to Ready, the voice appears in the Duaer Speech Voice list, and you can save it on a Duaer character.

Cloning is free. Speech is charged to the workflow owner's AI credits at the selected model's price. Refer to Duaer plans and credits.

Questions

Why is my cloned voice missing from the Duaer Speech list?

Duaer Speech lists only voices that are ready. Check the status under Settings › My voices: a voice in training needs a little longer, and a failed one needs a new recording.

What happens to Duaer Speech after I delete a voice?

In Duaer, Duaer Speech nodes and characters that use the voice fail until you pick another voice.

In this section

Nodes