An Open-source bot framework?

Yes, I partly agree. I think a good bot should be able to propose different "browsers" : one with Python requests module for example, which is fully customizable, and a "particular" option for Javascript eg a headless browser.

My bot framework is currently articulated around that.

ju111 said:

Yes, I partly agree. I think a good bot should be able to propose different "browsers" : one with Python requests module for example, which is fully customizable, and a "particular" option for Javascript eg a headless browser.

My bot framework is currently articulated around that.

Click to expand...

itz_styx said:

there is always a way i.e. if you heavily modify the code and randomize, but most people just use say selenium as-is. personally i write all my bots without headless browsers using direct HTTP requests.
i mean to make it easier if you aren't much into network programming you can use a headless browser, but for sites like IG and FB i wouldn't suggest that.

Click to expand...

Alright so HTTP requests might be what I should focus on since you dont have to worry about stuff like canvas spoofing etc

Im really not into network programming at all so I am trying to determine out of Selenium, Headless and HTTP requests which would be the best to start off on

Browser fingerprinting is pain in the ass. Some sites like Google can easily detect you. The real problem is you have to look like a real user which means you have to choose either Firefox 46 or Chrome 51 and reverse engineer it completely, then finally you make your own browser pool that can generate multiple browser profiles whenever you want and this isn't a job for one person.
There's already an app called Multilogin app which do this but I don't think it's open source. So you'll have to do it all manually.
One of the guys above mentioned he uses HTTP requests but I don't understand it. Hopefully he can shed light on us more.

Pain in the ass? More than that: a clear security failure.

Amazing that a remote website can list all browser extensions, all installed fonts, or the resolution of the screen. There is no obvious reason for this, and something wrong with the protocols. Mozilla should react, and propose better default configurations.

I'd like Go but as far as I know it is yet to have good browser automation but it has some projects for it. Any other aspect can be coverd, most if not all with std library. It has great concurrency model as well.
I wouldn't go for PHP since it has no real multy-threading (as well as Javascript with NodeJS which is one-threaded).
Python I don't like the language, in fact I do hate it.
I'd consider Java since it's got all what you need and most libraries and great tools for browser automation and logging and whatnot.

ju111 said:

No, not the same philosophy. For me, a click-and-play tools - on Windows - can only lack of flexibility about what can be done. What I need (a personnal need) is a flexible tool to code anything I want without coding multiple times the same features.

Click to expand...

You are right. But Browser Automation Studio supports code inserts in different language. AFAIK node.js is supported, other languages are to be added soon (but don't quote me on this)

What you are planning to do is cool but I would advise to conduct a proper market research first.

I'm interested in your open-source project. Keep me updated with a GitHub link, if applicable.