This is Experimental
VRChat/OSC by Speech Recognition
[Google Chrome]--(speech recognition)->[Python]--(OSC)->[VRChat]
- Python3.11 or Docker
- Google Chrome browser
- or other chromium can work? (not tested)
Spells are defined in config.toml
.
Please edit it at first.
make serve
make serve-docker
And open localhost:8080
.
There are 2 magic commands: /textchat/start
and /textchat/end
.
It starts/ends chat mode.
On chat mode, your spells will be sent as textchat.
Set spell.translate = true
in config.toml
and set ENV DEEPL_AUTH_KEY
.
[spell]
lang # Language for Chrome/SpeechRecognition
threshold # How rate you can allow for miss recognition
commands # An array of command, which has spell and OSC destination and arguments
translate # true or false; Enable to translate your textchat using DeepL
translate_langs = ["en", "zh", "ko"] # List of languages, which translate to
[vrchat]
send_port # 9000 by default
recv_port # 9001 by default
host # hostname or IP of VRChat. Usually "localhost" is OK.
A command has 3 fields.
{
spell, # Your speech text
dest, # OSC destination (e.g. /avatar/parameters/XXX)
args, # Usually, one parameter (int, float, string)
}
There are 2 special dest
, /textchat/start
and /textchat/end
.
Once /textchat/start
, later your speech will be broadcasted as textchat, until /textchat/end
.