Groovie

Vibing to Despacito
Our Website
About Groovie
Groovie's Features
Groovie's commands
GitHub Global Campus

Inspiration

After Youtube shutdown Groovy and Rhythm, in 2021 August, we struggled to find a cool music bot to vibe with on call. As most bots no longer risked accessing YouTube as they didn't want to get shut down too, which is where we had an epiphany. We realized that if we didn't monetize our bot we could access YouTube without any issues, so exactly what we did. We created Groovie in honor of Groovy but added a few features that we wished Groovy had, the ability to recognize what we say instead of needing to type it in chat using machine learning and vocal recognition.

What it does

At first glance, Groovie may appear like just another discord music bot, but in reality, this music bot has a very cool unique feature, it can actually comprehend what you are saying and process it to play music, pause, resume, and much more. This is all thanks to machine learning and voice recognition that allowed us to train it to recognize even the most agitated gamers' music requests.

How we built it

We made use of machine learning and voice recognition with discord.js and wit.ai, as well as a node.js package, called "discord-speech-to-text". After developing the base of the bot, we spent a few hours training the bot to sharpen its voice recognition skills.

Challenges we ran into

The biggest challenge that we ran into was training Groovie, to recognize its own name and a lot of other simple words like pause, resume and etc. In the beginning, Groovie kept thinking we were saying Ruby or Trudy or some other weird word, so we were forced to train it by reading articles, stories, and even each other's English essays to help develop a stronger model and better understand what we were saying. But thanks to the mixture of machine learning and a pre-developed vocal-recognition data model we were able to help Groovie be able to gain the comprehension abilities it needs to be a useful music bot.

Accomplishments that we're proud of

The biggest accomplishment that we're proud of was probably when Groovie played Despacito for the first time as it was proof of all our hard work finally coming together. So the first time we said, "play Despacito" and it worked we were ecstatic. All the training that it took to get there finally proved to be useful.

What we learned

We learned how to use machine learning and how although machine learning is powerful, the initial aspect of inputting data and making sure the bot interprets what we say accurately can be extremely tedious. We learned how to make a discord music bot using a variety of systems and methods.

What's next for Groovie

We will probably increase the features of Groovie, like being able to save playlists based on your voice, so that it has a database with playlists attached to your specific voice, adding a feature where the bot can determine what mood you are feeling by detecting your emotions through your voice tones and play a playlist based on the vibes your feeling. We also plan to give Groovie, it's own voice using text to speech (TTS) so that it can double as a chatbot that we can vibe with when lonely. Overall we are going to add a whole lot of new features to Groovie as we actually plan on using this music bot on a regular basis so we will continuously make edits and improvements to this music bot.