Recording Bot

A bot built to record and transcribe audio fragments from Discord.

Some functions taken from podbot.

Dependencies

• NodeJS
• Python3
• SpeechRecognition
• Discord.js
• node-opus

Usage

Install

• Set up the Cloud Speech Api and download credentials json.
• Create a Discord Bot.
• git clone https://github.com/sshh12/Recording-Bot.git
• npm install
• pip install -r requirements.txt

Run

• In main.js, change pythonapp to the name of one of the bot_*.py scripts
• Change token to the Discord bot token
• node main.js

Scripts

Depending on what you want to do with the recordings, you can select or write a python app and point main.js to it.

• bot_transcribe.py transcribes the speech of a user to their text channel.
• bot_repeat.py plays back each audio file.
• bot_conversation.py records conversations (prompt -> response) then replays response the next time it hears the same prompt. This requires create_conversation_log.py to preprocess audio files.
• bot_say.py plays lastest audio file with transcription of X when user says "say X".