erew123 / alltalk_tts

AllTalk is based on the Coqui TTS engine, similar to the Coqui_tts extension for Text generation webUI, however supports a variety of advanced features, such as a settings page, low VRAM support, DeepSpeed, narrator, model finetuning, custom models, wav file maintenance. It can also be used with 3rd Party software via JSON calls.
GNU Affero General Public License v3.0
1.09k stars 115 forks source link

Just wanted to say Alltalkbeta (v2) is absolutely amazing. I love it. Espescially with piperTTS! Comfyui Custom Node? #326

Closed FemBoxbrawl closed 1 month ago

FemBoxbrawl commented 2 months ago

I was just asking if you could possible create a custom node for comfyUI, which connects to the Alltalk address. (I am trying to get not only images but also text to speech, LLMs + much more all within comfyui, for custom AI workflows without having to constantly switch tabs, etc.

also, if you do add this, It would be cool to have different nodes that can conneect to the main node, such as an input text, as well as a text generated from an LLM, etc.? (not necessary even if nothing works)

erew123 commented 1 month ago

Hi @FemBoxbrawl

I did have a brief glance at ComfyUI's dev guide. Its certainly possible to do, though, it may be a decent bit of work. I've certainly never touched ComfyUI before as far as coding goes, so not sure how easy it would be to impliment.

What level of integration are you looking for? Just basic send text/use X voice? or more in depth?

FemBoxbrawl commented 1 month ago

Hi @FemBoxbrawl

I did have a brief glance at ComfyUI's dev guide. Its certainly possible to do, though, it may be a decent bit of work. I've certainly never touched ComfyUI before as far as coding goes, so not sure how easy it would be to impliment.

What level of integration are you looking for? Just basic send text/use X voice? or more in depth?

I think it would be nice to have the settings that are similar to the settings of Silly Tavern (but with Alltalkv2 (alltalkbeta) (with the the different tts models (xtts, SoVitts, piper, etc.) and also the RVC overlay option.

It is essentially just basics, but it would be also cool that you can add a basic text string (which can either be typed, or if used with a text generated by an LLM for example, to connect it to that alltalk node, and then it will use that text as TTS. but overall just basics integration. nothing too in depth.

erew123 commented 1 month ago

Hi @FemBoxbrawl

Sorry for the late reply, but I have very limited time at the moment. Im trying to crack through as many feature requests as possible, along with dealing with support etc, as well as balancing family matters.

I've added this to the Feature Requests here under the General heading. It will be one of those, when I get time, things. Its in the list though.

Thanks