A Discord bot designed to record voice chat audio, aimed at recording a specific podcast.
This bot has had plenty of problems actually doing what it's meant to for about a year or so now. Sometimes it records, sometimes it doesn't...
I've updated it to use all the latest in cunning fixes and workarounds but even now my testing has found it unreliable. My leading theory is that the boy backend needs to be close geographically to the server location.
I'm too short on time right now to look into this further alas. Ultimately the current official stance is that voice recording for bots is not supported at this point in time anyway.
Click here to add my instance of it.
It's worth noting that you'll need to host the bot yourself if you want access to the recording it makes - if you just use my instance you rely on me giving you the audio it generates.
Someone has actually adapted this project and wrapped it up in a user-friendly package.
Go take a look at podframe if you want to be able to just add a bot
and start recording.
(This is made by some chap by the name of _leanon over on this reddit thread)
Podframe is ded, you'll need to run your own podbot.
This bot has been tested against node.js 10.15.2
Make sure to use this version if you're having trouble getting it to work.
- Clone this repo and do the thing with the discord and the applications.
- Run
node index.jsornpm run podbotand go through the prompts to initialize configuration- when prompted enter your bot token (right click to paste into the terminal)
- the default options will set podbot up to be controllable by users with roles named "podhandler". Create this role on your server and assign it accordingly
- you can use the
controllers.usersarray to make specific users able to control the bot regardless of the server they're in
- If
./config.jsonalready exists this initialisation will be skipped. Edit it manually if you want to tweak it afterwards
- Update config.ini with your values (or pass in the env with whatever container orchestrator you're using)
docker build -t podbot .docker run -d --env-file config.ini --name podbot podbot- Your opus_string files are (by default) at /opt/podbot/podcasts, you will want to mount a volume there to get them out
- As of right now this container doesn't support processing the audio files
/podon- The bot joins the voice channel the command user is in and starts recording/podoff- The bot stops recording and leaves the voice channel the command user is in
The bot will generate audio fragments saved to podbot\podcasts\<channelId-timestamp>.opus_string.
These can be decoded and reassembled by running npm run process and picking the desired podcast session from the list.
If you've got the podcast recording folder in a non-standard location you can process it by running npm run process <path to podcast directory>.
This will generate a file for each recorded user with their discord ID as the filename. These files can then be imported into your favourite audio software (such as Audacity) side by side and everything should line up on the timeline nicely.
The output is signed 16-bit little endian PCM at 48KHz with 2 channels. Use the following options when importing into Audacity:
Once you've imported all the different user tracks you can then export in your audio format of choice.