[QUESTION] Bots and User-Agents ?

I have scripts that assign random user-agents to each account for each instance.
But I'm wondering If would be better to use one unique/specific user-agent string per account. And do not change it.

What's better ?

If you're experienced in this matter... any help would be appreciated !

When you log into X website in your browser as user Y, do you use a different browser each time? Do you change from mac to linux to windows to firefox to chrome to version 40 to version 13 to 26 all in a day?

No. Your bots shouldn't either.


But do the sites check? Probably not

It depends from site. More of them doesn't check useragent, they ofen check ips. But if you play with sites like google accounts they check useragents, ip and many more. Better for you is useragent permament for account

Using one specific user-agent string per account is best solution. In case of google if you change your last browser configuration it might ask for verification (like phone number etc).

With a bit of delay but, Many thanks for solving my doubt !

Make sense to have only one user agent per account. I made that change.

Thank you again.

Without any context it is impossible to answer. Obviously it depends on the website. If you want to bypass cloudflare for example, you have to use the same user agent on every request for that particular IP. But this is only one specific example.

mantic0re said:

Without any context it is impossible to answer. Obviously it depends on the website. If you want to bypass cloudflare for example, you have to use the same user agent on every request for that particular IP. But this is only one specific example.

Click to expand...

I'm mainly using Tumblr. And since Rootjazz owns a bot for that social network I though that I could trust him/her. Hehe

Well for tumblr and other big, "wide-net" sites you can be certain they are cataloging your devices associated with an account. People will hop on with their desktops sure, but they also chill on the couch with a laptop/tablet(both?) and pick up a phone when its more convenient or they are not at home.

So you aren't going to necessarily get flagged for changing user-agents, but what Rootjazz describes is activity that is flat out unnatural and that is kind of the point he was making. Bots trying to "secretly" mimic a human should not do bizarre things.

I wouldn't use truly unique user agents for your bots, but some variety in genuine user agents may be a good way to avoid giving them a foothold to identify you as a bot user. EG: "user agent A + behavioral statistic C = definite bot".