Instapy: Making the bot harder to identify

[AlainPilon]

I think we should randomize all the numeric values used in the app (wait time, number of likes, comments, etc..) using a normal distribution. This would be more realistic of human behaviour.

Any monitoring system would find it odd that the user is always linking exactly 50 images and always waiting 2 seconds between clicks...

My idea would be to wrap any number into a lambda which would return the randomly normalized number.

Or am I too paranoid?

[timgrossmann]

I did a lot of the randomness with cron and the quickstart.py file. But you're right, this should probably be part of the tool

[pekapa]

I'm up for doing that as I was thinkg about it just yesterday.

[ToJen]

@pekapa same here

[AlainPilon]

same here... Push something and we can share ideas.

[timgrossmann]

@pekapa @ToJen @AlainPilon Feel free to make a PR with your changes.

[pekapa]

Just pushed #100 to help with this.

[eduardocalazansjr]

As i said, instagramweb seems to gather all the mouse/keyboard events via javascript.

I think that simulating the page movement is more efficient than random sleep times (also with the page loading times, it will be equivalent to a random sleep time).

Random sleep times without some page interaction will still flag as a bot.

Edit: Selenium has nice mouse movement features, using with random x, y values while doing the movement to the button will result in random times on pages.

ref

[stale[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

If this problem still occurs, please open a new issue