A discord bot (Bob Assistant) with ASR, GPT, play music from Youtube, and more...
- A Bot on Discord Developer Portal
- ffmpeg python3 python3-venv python3-pip libtool espeak
- nodejs
- pip: openai-whisper TTS
- npm: install (use package.json)
- .env: BOT_TOKEN, APP_ID, GUILD_ID, YOUTUBE_TOKEN, TEXT_CHANNEL_ID, VOICE_CHANNEL_ID
First, you might want to create .env file with the keys above:
- BOT_TOKEN: The Token of the Discord Bot (check out the Discord Developer Portal)
- APP_ID: OAuth2 Client ID of the Bot (check out the Discord Developer Portal)
- GUILD_ID: Server ID (you should turn on Developer mode on Discord in order to see this)
- YOUTUBE_TOKEN: Token of Youtube Data API
- TEXT_CHANNEL_ID: Default Discord text channel for Bot to send text response in
- VOICE_CHANNEL_ID: Default Discord voice channel to join in automatically whenever bot is online
(Recommended) Using Docker, build image, then run container:
docker build -t discordbob:v1.0 .
docker run -d -it --restart always --env-file .env --name DiscordBobGPU --gpus all discordbob:v1.0
Or if you have installed everything manually, run:
node .
The bot has 2 states:
- Listening to wake-word state (state 1)
- Listening to command state (state 2)
In state 1, only voice line with Hey, Bob will be accepted and then switch to stage 2.
In stage 2, there are some voice commands below:
- ['play music', 'play the music', 'play song', 'play the song', 'play track', 'play the track', 'phát nhạc'] : to play music from Youtube search
- ['switch language', 'đổi ngôn ngữ'] : to switch the ASR language (in case you want to play songs in another language. I added English and Vietnamese only. You can add more if you want)
- ['reset bot'] : reinitialize the bot, reset its configurations and state to default.
You can also deploy your slash commands to your Discord server by using:
node deploy-commands.js
And then use the slash commands in your Discord server.