Inspiration
We wanted to create an Artificial Intelligence from scratch, and the idea of giving our own flair to something existing sounded very cool. Inspired by this idea, we proceeded with our WordleBot.
What it does
Solves Wordle in the minimum amount of guesses aided by our algorithm without taking up excessive time.
How we built it
We used Python to build the model and Javascript, React, and Tailwind to create the website. We also used Selenium to make the model interact with Wordle's official website and solve puzzles flawlessly.
Challenges we ran into
We started off with an initial model which would solve the puzzle for a lot of problems but would get stuck/take a long time/guess wrong. However, we slowly optimized the model to be significantly more accurate and efficient. Finally, we optimized the model to be faster, and take less time while running while retaining its accuracy. We have also trained the model to learn about the best guesses to take based on situations, enabling it to make good first guesses based on experience.
Accomplishments that we're proud of
We took an idea that was cool to us to a fully working model in just a few hours. We worked with a lot of new languages that we were either completely new to or weren't fully comfortable with.
What we learned
We learned a lot about Selenium, Tailwind, JavaScript, React, and project management in general.
What's next for Wordle Bot
We plan to make the bot able to complete around 2000 puzzles all by itself with 100% accuracy all while taking the smallest amount of time possible. We also want to implement our refined and more accurate bot into the main game to offer a more challenging and competitive Wordle experience.
Log in or sign up for Devpost to join the conversation.