A discord bot that records a voice channel and has the ability to upload the clip to an AWS S3 bucket
install java 8+
install ffmpeg https://www.ffmpeg.org/
gradlew build
Modify application.properties with your bot token and a path to save clips. If using an AWS S3 bucket, it must have public access allowing reads and writes (If any s3 experts would like to point out how I can enhance this to disallow public writes, I am interested in learning).
# your discord bot token from https://discordapp.com/developers/
botToken=
# character that tells the bot to listen to a command
commandCharacter=!
# local filesystem path that the bot should save clips to
recordingPath=
# the format that the raw output should be converted to when ending a recording
recordingFormat=flac
# Should the bot attempt to use the AWS feature
# if false the upload command will attempt to upload the file to the discord channel directly
# discord file size limits aren't very large, so long recordings will likely fail to upload
useAWS=false
# The AWS bucket name that the file should be uploaded to
awsBucket=gradlew run
The bot listens for chat commands in your server. When a command to start recording is received the bot will join the voice channel that you are in and start recording.
!start starts recording
!end ends recording and endcodes the raw data into the configured output format, bot will remain in the voice channel until encoding is finished.
!upload uploads the most recent encoded file to discord, or the configured AWS S3 bucket.
Q: Why is does the bot appear as if it is talking when it is recording?
A: The audio recording API is undocumented and unsupported as far as I can tell. To be able to receive audio you must also be sending audio at the same time. To workaround this, the bot sends silence while recording.