patrickjs/auto-maple

Auto Maple is a program that plays MapleStory, a popular online role-playing game, automatically on your behalf. Instead of you manually clicking and pressing keys, the program watches your character's position on screen and controls its movements and attacks to grind through game content hands-free. The program works by combining a few clever computer vision techniques. It watches the game's minimap (the small position indicator in the corner) to track exactly where your character is at any moment, then uses a pathfinding algorithm to figure out the best route to reach objectives. It can also recognize and solve in-game arrow-key puzzles automatically by analyzing what it sees on screen. When it comes time to attack, it presses the right combination of ability keys in the correct sequence, imagine recording a specific combo move and having it replay perfectly every single time. The system is built to be flexible. Instead of hardcoding every character class and playstyle, you provide two things: a "command book" (a Python file that lists what keys each ability should press) and a "routine" (a CSV spreadsheet that tells the program where to move and which abilities to use at each location). This modularity means one person can set up a command book for a Mage, someone else creates a routine for a specific dungeon, and the community can share these resources so everyone benefits. The routine system even supports loops and conditional logic, so complex farming patterns can be encoded. This kind of automation is primarily used by players who want to grind repetitive content without sitting at their computer for hours, or by the community members who enjoy the puzzle of optimizing strategies and sharing their setups. The project requires some technical setup, installing Python, CUDA for GPU acceleration, and downloading a pre-trained machine-learning model, but once configured, users just select a routine and hit go.