ComfyUI Node
VibeVoice TTS
A ComfyUI node in audio/tts with 16 inputs and 1 output.
VibeVoice TTS
- speaker_1_voice
- speaker_2_voice
- speaker_3_voice
- speaker_4_voice
- AUDIO
◄model_name▾►
◄text[1] Hello, this is a cloned voice.
[2] And this is a generated voice, how cool is that?►
◄quantize_llm_4bitfalse►
◄attention_modesdpa►
◄cfg_scale1.30►
◄inference_steps10►
◄seed42►
◄do_sampletrue►
◄temperature0.95►
◄top_p0.95►
◄top_k0►
◄force_offloadfalse►
Categoryaudio/tts
Inputs (16)
| Name | Type | Default | Description |
|---|---|---|---|
| model_name | COMBO | Select the VibeVoice model to use. Official models will be downloaded automatically. | |
| text | STRING | [1] Hello, this is a cloned voice. [2] And this is a generated voice, how cool is that? | The script for generation. Use '[1]' or 'Speaker 1:' for speakers. If a speaker in the script lacks a reference voice, it will be generated via zero-shot TTS. |
| quantize_llm_4bit | BOOLEAN | false | Quantize the Qwen2.5 LLM to 4-bit NF4 via bitsandbytes. Diffusion head stays BF16/FP32. |
| attention_mode | COMBO | sdpa | Attention implementation: Eager (safest), SDPA (balanced), Flash Attention 2 (fastest), Sage (quantized) |
| cfg_scale | FLOAT | 1.300.1–50 | Classifier-Free Guidance scale. Higher values increase adherence to the voice prompt but may reduce naturalness. Recommended: 1.3 |
| inference_steps | INT | 101–500 | Number of diffusion steps for audio generation. More steps can improve quality but take longer. Recommended: 10 |
| seed | INT | 420–18446744073709550000 | Seed for reproducibility. Set to 0 for a random seed on each run. |
| do_sample | BOOLEAN | true | Enable to use sampling methods (like temperature and top_p) for more varied output. Disable for deterministic (greedy) decoding. |
| temperature | FLOAT | 0.950–2 | Controls randomness. Higher values make the output more random and creative, while lower values make it more focused and deterministic. Active only if 'do_sample' is enabled. |
| top_p | FLOAT | 0.950–1 | Nucleus sampling (Top-P). The model samples from the smallest set of tokens whose cumulative probability exceeds this value. Active only if 'do_sample' is enabled. |
| top_k | INT | 00–500 | Top-K sampling. Restricts sampling to the K most likely next tokens. Set to 0 to disable. Active only if 'do_sample' is enabled. |
| force_offload | BOOLEAN | false | Force model to be offloaded from VRAM after generation. Useful to free up memory between generations but may slow down subsequent runs. |
| speaker_1_voiceopt | AUDIO | Reference audio for 'Speaker 1' or '[1]' in the script. | |
| speaker_2_voiceopt | AUDIO | Reference audio for 'Speaker 2' or '[2]' in the script. | |
| speaker_3_voiceopt | AUDIO | Reference audio for 'Speaker 3' or '[3]' in the script. | |
| speaker_4_voiceopt | AUDIO | Reference audio for 'Speaker 4' or '[4]' in the script. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| AUDIO | AUDIO | — |