Against the TOS to create bot to respond to @-mentions?


Hello, sorry that I didn’t check the TOS myself. I just had this thought and figured it’d be easier to parse for someone with the mental model already stored. I’m sorry if this is lazy, but promise I’ll pay it forward. :slight_smile:

(Plus I finally wanted to post here!)

My Twitter handle is @echelon, and it just happens to be a term that Jared Leto / 30 Seconds to Mars (a music band) fans use to self-identify within their fanbase. At least that is my understanding of how they use the term.

For the most part, nothing bad has come from this domain overlap aside from the deluge of mentions I receive from Jared Leto fans. I would have no problem with the message volume if it were not for the mental filtering I have to do in order to find valid/relevant mentions. I suppose I could ignore all Jared Leto-related mentions, but this has also had the side effect of crowding my inbox; I had to disable notifications because of it. The simple fix would be to just create another account, but I don’t want to change my handle since I’m echelon everywhere else (and have been since the early-00’s).

That brings me to the point of my inquiry: would it be against the TOS to create a bot to heuristically respond to @-mentions that exceed some weighted “Jared Leto” threshold criteria?

Say that I can algorithmically determine that a tweet is related to Jared Leto or the band 30 Seconds to Mars based on machine learning or on simple keyword matching. I would like to respond with a single message (only once per handle no matter how far spaced the messages are) that kindly directs them to more relevant Twitter users, such as Jared Leto’s official account. I would only send one reply to any given user. (Once replied, their name would be appended to a list never to automatically respond to again.)

Is that against the TOS? I fully understand and respect the terms if it is. I was just curious and thought the idea would be a fun mental exercise.

If the above behavior is currently permitted, would it still be acceptable if I were to respond instead from a second account, such as the purely-hypothetical “@echelonIsNotJarded”? My desire to use a second account would be so that I don’t spam the few actual followers I do have; I think they’d be annoyed by seeing dozens of automated messages being tweeted from my account. If this were not permitted, I also totally understand. It seems like it could lead to abuses or annoyance.

I’m probably just going to have to think up another strategy altogether.

Is there an API layer that sits between Twitter notification and the incoming tweet stream? I’d love to flag incoming tweets as not deserving a push notify (or filter it out altogether).

Sorry for the long winded post. Thanks in advance to anyone taking the time to respond! :slight_smile:


This is good question. First thing I’d point you to is our Automation Rules and Best Practices, which describes the principles behind how we view automation. What you describe (responding to people @-mentioning your handle) is just within the guidelines—and making it once per user makes the opt out provision moot.

HOWEVER, you may yet find yourself suspended if recipients of your messages decide that they’re spammy, and report them as such.

Responding from a second account is a non-starter. The automated replies in that case would look unsolicited, and our own spam detection systems don’t take kindly to that sort of thing—as well as these replies being vastly more likely to be reported as spam by the people receiving them.

Sadly there’s no API layer or plugin mechanism for you to determine which Tweets should get pushed to you. Neat idea, though.


Thanks so much, Isaac!

I’d rather not skirt the fine line on this given the triviality of the use case. It’s great to get an official answer. :slight_smile: