Don't Ban Us

Avoid getting banned on your favorite games while still talkin' trash!

Comment

Inspiration

When you find a game you enjoy, the last thing you want is to be banned from it. Don't Ban Us is a tool users can leverage to make sure their language is always appropriate to their opponents and never is the reason why they get blocked! We built this project from our shared experience of playing games and realizing the importance of fair and respectful interactions within the gaming community.

Our mission is to promote a positive gaming environment where players can enjoy their favorite games without fear of being banned due to offensive language or behavior. With Don't Ban Us, we aim to foster a sense of camaraderie among gamers while upholding the principles of fair play and sportsmanship.

What it does

Don't Ban Us is a plugin that gamers can install on their machines and apply to games of any shape and size. When users open the chat in a game and type a rude or toxic message, Don't Ban Us will replace their text with more game-friendly phrases that still insult their opponents. Roast other people in the chat without fear of losing game privileges! Don't Ban Us is supportive of positive chat messages and allows those to pass through unfiltered.

Before: My grandmother plays League better than you, you stupid loser.

After: Your League skills must come from your grandmother, because they sure don't come from you!

How we built it

Don't Ban Us is built atop Electron. With powerful features like built-in macros, system tray, and end-to-end connection, it gave clarity to the road ahead. When users press certain inputs, we intercept these and process them through functions, then use a robot-type feature that lets us override the text users have in their chat boxes.

The backend is built with Python! We used the nltk (Natural Language Toolkit) for sentiment analysis on user messages to detect Negative phrases. Once a message is determined to be rude, we use GPT's API to rephrase the sentence into kinder yet still insulting text.

Challenges we ran into

We ran into some recursive issues with registering key binds! For example, we had some cases where we needed to register keys to trigger our backend script, unregister them so we could use those keys to type text, and then reregister them again to be ready for the next time a user clicked it. We also experienced challenges with connecting Electron to our Python backend as it's not very well documented. Since we are building Don't Ban Us for fast-moving game environments, we realized the importance of optimizing our tool's performance so we worked to save trained models and quicken our GPT calls.

Accomplishments that we're proud of

We were proud of the first time that we saw our tool run end-to-end! It was exciting to watch the backend calls get triggered after we typed an insulting message and see the automated typing complete as the message was replaced. We were very satisfied to have successfully overcome our merge conflicts and connect the backend and frontend of our project!

What we learned

Our team has been able to try out working with new frameworks and libraries that we haven't experienced before, from the GPT API, to nltk, to Electron.

What's next for Don't Ban Us

We envision Don't Ban Us to be the future of standardized chat moderation across platforms ranging from gameplay to social media. Currently, it censors rude/profane chat messages but we have explored extensions of the tool to filter sensitive information (think credit card numbers, addresses, phone numbers, and ages) and adapt to prevent racial slurs on popular platforms. Our goal is to provide chat moderation features to games of all sizes (established ones with large development teams as well as smaller projects with fewer technical resources).

Built With

Share this project:

Updates