I have a bot that sends out whispers to users when a channel begins streaming. The bot only messages people that explicitly want to get informed.
The bot stopped sending out whispers on 17-02-14 so my first thought was it is related to this change, because it was as the exact same date: https://blog.twitch.tv/action-required-twitch-api-version-update-f3a21e6c3410#.k23opfo49
However, if I use my existing code and use another account + oauth token then the whispers are working.
However, I can still send whispers when I log in with the bot account on the twitch page and write myself a message.
So how do I find out whether the bot account got banned from sending whispers or if I have to fix the code? Is there a difference in sending whispers via IRC and via the Twitch website?
Have you tried using your code with a different bot account?
Seems like that’d be a sure fire way to know wether your code still works.
Also, whisper bans are usually like chat spam timeouts, they get lifted after a period of time.
Yes, I tried the (unchanged) code with another bot account and it was working as intended.
If it is only just a timeout than it sure is a long one, since it has been over a week now.
There isn’t a difference in code at all. This sounds like you were shadowbanned for sending too many messages or too many reports of spam by users. You can apply for whitelisting for whispers via support, but there isn’t a guarantee of whitelisting.
I looked through the whispers, there definitely wasn’t an abuse by someone else. I was sending about 10 messages 2-3 times a week (only to users who opted in voluntary) and each message clearly states how to unsubscribe from the feature so I don’t think it received much spam reports.
But I realize that sending the same message to several people at the same time over and over again can easily be interpreted as spam.
Thanks a lot for the reply, I will look into that.
There isn’t a concept of “opt-in” with our Whispers system and how it detects spam. It’s a combination of frequency of messages, number of recipients, and reports from users. Some users sign up for these things and forget, and then they flag as spam. 
Well the support told me that they are not whitelisting at the moment…
@marci4 We’re still whitelisting. You likely do not meet the criteria to be whitelisted. Our internal criteria center around the bot being a proven good actor, number and quality of messages sent, and security of the bot account.
This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.