Hello, Reddit!
Voice chat features of various AI-Services
including character.ai but also SillyTavern's own Extensions itself
have always bothered me, because they do not run truly hands free.
an extremely big annoyance of character.ai's version
was that it REQUIRED the user, to talk, for the AI to generate the next message.
This is NOT how communication works.
People pause. People breathe.
Sometimes you literally have nothing to say to a reply.
This is simply unacceptable User Experience.
While researching, if something like this existed already in the SillyTavern Extensions found online.
I found a barely maintained repo, which I have then forked intending to do a simple fix.
Unfortunately, also this Extension lacked the features of a Truly Hands Free Chat Experience.
So it had to be rewritten.
and now the extension (to my knowledge) works exactly as I have imagined.
Behold, what character.ai wished their call mode was capable of! 😉
Introducing the (to my knowledge) first,
simple to setup Hands-Free-Voice extension
in the Advanced Roleplaying AI Scene.
It turns SillyTavern into a proper voice call experience:
- Character finishes speaking (real audio end detection)
- Mic opens automatically
- You speak naturally (pauses are respected)
- Whisper transcribes (Groq / OpenRouter / local)
- Your message is sent + character replies
- If you stay silent → it auto-continues and the character replies.
No push-to-talk. No keyboard. Just talk, Hands free.
**Features:**
- Full Auto (no forcing you to say anything to get a reply)
- Configurable Timeout + Reply pause tolerance + max recording length
- Optional quote wrapping
- Works together with the Default SillyTavern TTS Extension
Repo + full instructions: https://github.com/Flaxify/ST-Hands-Free-Voice
Tested on the latest SillyTavern 1.17.0.
Using Whisper via OpenRouter: https://openrouter.ai/openai/whisper-large-v3-turbo
Requires working TTS + API key to a Whisper provider (Groq / OpenRouter / local)
Would love feedback!
~Thomas