KittenTTS Node for Voice Generation

KittenTTS Node for Voice Generation

This Python class defines a custom node KittenTTS intended for generating audio from text using a selection of predefined voices. It is categorized under "utils" and integrates with a node-based workflow system.

---

Description

• The KittenTTS node accepts a text string and a voice selection from a fixed list of available voice models.
• It calls the generate_audio function (imported from the local .generate_voice module) to produce synthesized speech audio.
• The node outputs the generated audio data, making it usable in downstream audio processing or playback nodes.

---

Input Parameters

• text (STRING):
  The input text that will be converted to speech.

• voice (STRING):
  The voice model to use for synthesis. Allowed values are:

  • expr-voice-2-m
  • expr-voice-2-f
  • expr-voice-3-m
  • expr-voice-3-f
  • expr-voice-4-m
  • expr-voice-4-f
  • expr-voice-5-m
  • expr-voice-5-f

---

Return Value

• Returns a tuple containing one item of type AUDIO — the synthesized speech audio.

---

Integration Details

• INPUT_TYPES class method defines the input interface, listing required parameters and valid voice options.
• RETURN_TYPES specifies the output type as "AUDIO".
• FUNCTION names the method apply_generate_voice to be called when executing this node.
• CATEGORY groups this node under "utils".
• NODE_CLASS_MAPPINGS links the identifier "KittenTTS" to the KittenTTS class.
• NODE_DISPLAY_NAME_MAPPINGS sets the display name as "Generate voice" for UI presentation.

---

Usage Example

kitten_tts_node = KittenTTS()
audio_output, = kitten_tts_node.apply_generate_voice(
    text="Hello, how are you today?",
    voice="expr-voice-3-f"
)
# `audio_output` contains the generated audio data, ready for playback or saving.