Extensions/Audio8
ComfyUI Extension

Audio8

Audio8 Text to Speech / Speech to Text nodes

By niknah·Created about 18 hours ago·Updated about 14 hours ago· 0
niknah/ComfyUI-Audio8
Nodes—
On cloudLocal install
Stars0
Updatedabout 14 hours ago
Readme

ComfyUI nodes for https://github.com/Edge0-AI/Audio8_TTS/

  • Text to speech (Edge0/ARK-ASR-3B, Edge0/ARK-ASR-0.6B)
  • Speech to text (Edge0/Audio8-TTS-Realistic-Air)

Example workflows or you can get them from the templates section in ComfyUI

Note: The models use trust_remote_code=True, which means whomever controls the huggingface repo can do things on your machinhe. Same with this extension. If you'd like to use .safetensors, try this extension for Audio8-TTS-Realistic-0.6b TTS: https://github.com/Pepehoschi/ComfyUI-Audio8TTS/

| Model | VRAM | | :--- | ---: | | Audio8-TTS-Realistic-Air (uses more with multiple voices) | 2GB+ | | ARK-ASR-0.6B | 3GB | | ARK-ASR-3B | 8GB |

Multi voices output...

Put your sample voice files into the input folder like...

LanguageTest.flac
LanguageTest.txt
LanguageTest.French.flac
LanguageTest.French.txt
LanguageTest.German.flac
LanguageTest.German.txt

Then you can use prompts for different voices...

{main} Could you tell us where the best place for fish & chips is around here?
{German} You must try the little place down by the harbour, just past the ferry wharf, because the fish is caught fresh that morning.
{French} Perfect, let’s grab lunch there first!

The main voice refers to the one with no name LanguageTest.flac.