Voice Generation
The Voice Generation node generates spoken audio from text, using a voice or model you choose. Use it for narration, dialogue, or voiceovers.
Inputs and outputs
Inputs: Text (always), and Voice Ref (an audio clip to clone or reference), shown only for models that support voice cloning. Plus a generic Flow input.
Output: a single Audio handle.
Writing text
Type the text you want spoken into the box at the bottom of the card ("Enter text to generate speech..."). If Text is connected from elsewhere, this box is disabled and shows "Text connected from input" instead.
The generated audio plays in a well at the top of the card. Click Download in the settings bar to save it.
Choosing a voice and model
Click the model picker to search and choose a model. Models that ship preset voices show a Voice selector directly on the card. Models built for cloning hide this selector, since the voice comes from your uploaded reference clip instead.
The default model is Chatterbox TTS. This node always follows your account's default provider (fal.ai, OpenRouter, or Kie AI); it doesn't have its own per-node provider override.
fal.ai's curated list includes Chatterbox TTS (voice cloning), ElevenLabs Turbo v2.5 (22 preset voices), MiniMax Speech 2.8 HD and Turbo (17 preset voices each), Seed Speech v2 (41 preset voices), Grok Voice TTS (5 voices), Gemini 3.1 Flash TTS, Inworld TTS-1.5 Max (113 preset voices), Qwen3-TTS, Index TTS 2.0 (voice cloning only), VibeVoice 7B, and MiniMax Voice Clone (voice cloning only). OpenRouter and Kie AI offer a further set of models, including ElevenLabs models on Kie AI.
Voice Settings & Audio Cleanup
Click the sliders icon in the settings bar to open Voice Settings & Audio Cleanup. What you see depends on the model:
- Delivery: a Style prompt and/or an Emotion prompt, for models that support them.
- Voice Settings: sliders that vary by model, such as Exaggeration, Temperature, and CFG Scale for cloning models, Speed for models that support it, or Stability, Similarity, and Style for ElevenLabs-style models.
Click Reset to defaults to put every slider back to its starting values.
Running the node
Select the node to reveal its Run button. A standalone node with no inputs runs itself; run it again to also run everything downstream. If its inputs already have results, running it runs everything downstream too. If inputs aren't ready yet, you'll be asked whether to run the whole workflow, or run everything up to and including this node.
Click the Inputs / Outputs button at any time to see exactly what data went in and came out.