I’m a programmer and looking to join forces with some people to perform some social studies online, as a user since the inception of Twitch, the area that I would love to tackle is twitch. Specifically, using an NLTK and chat logs to analyse some user behavior on Twitch.
I dont have any questions about the API as it is well documented, the obvious first thing to do is assess whether it is allowed. Botting on Twitch is very common place, and there are even popular services like rechat.org that appear to log all chat and offer a non-commercial plugin to review chat logs while viewing streams as a nifty little idea.
However, from my initial assessment, it seems this is outright forbidden in the TOS, source:
( http://www.twitch.tv/user/legal, “Twitch Terms of Service - Last modified on 1/15/2015”, Accessed, 01/08/2015)
The Twitch Services may include interactive areas or services (" Interactive Areas “), such as chat boxes or web forums, in which you or other users may create, post or store content, messages, materials, data, information, text, music, sound, photos, video, graphics, applications, code or other items or materials on the Twitch Services (” User Content " and collectively with Broadcaster Content, " Content ")…
You agree that you will comply with these Terms of Service and Rules of Conduct and will not:…
…use any robot, spider, scraper, crawler or other automated means to access the Twitch Service…
So here we see that chat falls into the defined term “Twitch Service”, and any form of crawling/scraping such as botting or a spider is not allowed.
I’ve looked extensively and do not see any legal remedy to this where they later make an express legal claim that the prohibition is excepted under X or Y circumstance.
So, as somebody whos started using Twitch around the same time as Justin.tv and own3d.tv, (the start), obviously I think this falls in flat contrast to the actual user behavior. I imagine if any moderators/administrators ruled on this policy, they would say this exists as a preventative measure, and perhaps a sensible implementation of the rule is just to ask individual channels for permission before doing so. At least, this makes sense in when you consider it against the fact that every large channel on twitch uses custom botting to make the experience better for their audience.
Anyway, if anybody twitch admins can refer me to a policy that supercedes this one, if any user can link me to a public thread that states an exception to this policy, or if an actual employee of twitch could provide me any ruling on this, i would really appreciate it.
I’m only being so formal about wanting something in writing, because I dont want to invest a lot of time/money then get a gag order about using information I obtained that is in clear violation of their policy (even though there appears to be rampant violation of said policy). Also, sorry if this has already been answered, I did some searching and did not run into a similar thread, and I figured making a public discussion on this part of the TOS might be good for the API community.