ComfyUI Node
Miso TTS - Generate
Generate ComfyUI AUDIO from text, with optional prompt audio/reference context.
Miso TTS - Generate
- miso_model
- reference_audio
- audio
◄textHello! This is Miso TTS running inside ComfyUI.►
◄reference_text►
◄speaker0►
◄max_audio_length_seconds10.0►
◄longform_chunkingfalse►
◄words_per_chunk80►
◄temperature0.90►
◄top_k50►
◄seed0►
CategoryMiso TTS
Inputs (11)
| Name | Type | Default | Description |
|---|---|---|---|
| miso_model | MISO_MODEL | Output from Miso TTS - Load Model. | |
| text | STRING | Hello! This is Miso TTS running inside ComfyUI. | Text to synthesize. |
| reference_text | STRING | Transcript for reference_audio. Connect Miso TTS - Whisper Transcribe here. | |
| speaker | INT | 00–99 | Speaker tag used in the prompt, e.g. [0]. Leave 0 for normal single-speaker use; change only for multi-speaker context. |
| max_audio_length_seconds | FLOAT | 10.00.08–120 | Maximum generated audio per chunk in seconds. |
| longform_chunking | BOOLEAN | false | Split long text at sentence boundaries. Each chunk reuses the original reference audio if connected. |
| words_per_chunk | INT | 800–500 | Target words per longform chunk. Longform may lower this to fit max_audio_length_seconds. 0 disables text splitting. |
| temperature | FLOAT | 0.900–2 | Sampling temperature. 0 is greedy. |
| top_k | INT | 501–2051 | Top-k sampling over the 2051 audio vocabulary. |
| seed | INT | 00–2147483647 | Random seed. 0 picks a random seed. |
| reference_audioopt | AUDIO | Optional prompt audio for voice/reference context. This is not guaranteed speaker matching. |
Outputs (1)
| Name | Type | Description |
|---|---|---|
| audio | AUDIO | — |