There is no difference when it comes to packet manipulation between a proxy and a clientless bot. Both have direct access to the packets. The difference between the two just comes down to how much work you want the program do to. A proxy only has to relay packets between the client and the server, meaning the client is still doing the work of figuring out what needs to be sent and when. And of course a proxy can also send its own packets on behalf of the client if it wants, which is what the plugins for k-relay are doing. But if you look at a clientless bot, you no longer have the client dictating what and when packets are going out, you now have to send all the packets that the client would normally be sending. The advantage of a clientless program over a proxy though is the cpu & memory usage on your system. If you run a proxy, you have the proxy's usages + the clients usage (which flash is a horrible memory hog). And the reason you see people making proxies instead of modifying the actual game client just comes down to C# being easier to program in the actionscript haha.
I also wanted to make an update that i tested this with 3 accounts and then 4 accounts. It seems the memory and cpu usages of the program sticks around 0.5% - 2%, which is amazing in my opinion. It has an initial spike where it hits like 8%, but then drops down and stays around 1% pretty much. Compared to the game client which for me runs 30%+ just for 1 account, this is pretty cool to me.
Edit: wanted to post that when i compiled the program in release mode instead of debug, and i had 19 accounts running, the cpu was actually about 1-2% lol (19 clients in debug mode ran around 7-10% for me). Running the program in release mode removes a lot of the debugging crap that is added to the debug mode version, im actually shocked at how little memory it uses for 19 clients.

I also wanted to make an update that i tested this with 3 accounts and then 4 accounts. It seems the memory and cpu usages of the program sticks around 0.5% - 2%, which is amazing in my opinion. It has an initial spike where it hits like 8%, but then drops down and stays around 1% pretty much. Compared to the game client which for me runs 30%+ just for 1 account, this is pretty cool to me.
Edit: wanted to post that when i compiled the program in release mode instead of debug, and i had 19 accounts running, the cpu was actually about 1-2% lol (19 clients in debug mode ran around 7-10% for me). Running the program in release mode removes a lot of the debugging crap that is added to the debug mode version, im actually shocked at how little memory it uses for 19 clients.


