Wordle Solver

How it works

Essentially this solver uses Puppeteer to open the web page and repeatedly type guesses and eliminate potential solutions from a pool. Imagine divide-and-conquer, just way more convoluted than necessary. After trying any given word, the program keeps track of which letters are present, absent and those that are already in the correct place - this data is used to filter down the dictionary and inform the next guess. Currently, there's no additional heuristic selection to choose the next word from a list of candidates - simply put it's entirely unnecessary (most words will be solvable somewhere in the order of 10 guesses). Do note that if you intend on running this (why?!) you'll need >=Node 16 as I make use of ESM.

Difficulties

Turns out nested Shadow DOMs make things hard - read the source code at your own peril, you were warned.

Learnings

In all seriousness, building this solver has genuinely proven a valuable exercise - I've never touched web scraping/emulation before and, despite being thrown in the deep end, I'd say it's gone pretty well.