Rap battle-based social media
https://drive.google.com/file/d/1kUrwjqG8MmXdljeB9AAHqtB4A0qWfbXM/view?usp=sharing
Our inspiration for creating this project was the belief that an epic rap battle between celebrities and historical figures would be a novel and entertaining concept that could bring together audiences from all walks of life. The idea of seeing our favorite personalities and iconic figures engaging in a rap battle was too good to pass up, and it had the potential to be a truly engaging and memorable experience. (unites people)
Moreover, we were fascinated by the idea of using technology to facilitate such an undertaking. Utilizing advanced natural language processing algorithms, stable diffusion models, and other cutting-edge technologies would be essential for generating high-quality rap lyrics and producing a seamless and immersive experience for the user.
We also recognized that the potential combinations and matchups were endless, offering hours of entertainment and creative expression for participants and viewers. From Joe Biden and Obama to Drake and Elon Musk, the possibilities were genuinely limitless.
Our project, SpitFire, is an application that enables users to create and experience realistic and authentic rap battles between their favorite celebrities and historical figures. With the option to select from many rappers, users can generate struggle between two sides that reflects their personal preferences and interests.
One of the critical features of SpitFire is the ability for users to enter a specific topic or theme, which can be used to tailor a rap battle to the user's particular needs. This allows for a personalized and engaging experience that is true to the personalities and lives of the actual rappers themselves.
The use of advanced natural language processing algorithms and machine learning techniques ensures that the generated lyrics are of the highest quality and reflect each rapper's unique style and cadence. The result is a truly immersive and entertaining experience that brings the art of rap to life in a new and exciting way.
We built our app using React.js, a modern and versatile JavaScript framework. Regarding the front-end development that took place, it was developed using HTML, CSS, Node.js, and Styled Components. Most of the user interface (UI) was done by Styled Components, allowing us to build the visual aspect of our application quickly. In order to create the technical and functional aspect of our application, we first utilized GPT-4 to generate rap verses for each artist in each rap battle. In order to convert the text to speech, we used the text-to-speech tool from elevenlabs to create the audio files for each of the rap verses. Now that the rapping was done, we now needed to create animations of the rappers rapping back to each other when it was their verse. To do this, we created animations using a Stable Diffusion model from Replicate that would create short videos of the “rappers” mimicking speaking. Afterwards, we used Python algorithms to prepare the final video. Solidity was used for the creation of smart contracts. We also developed a Next.js full-stack application for the frontend and backend of our platform. Finally, we created an API with Python Flask in order to deploy the final video and connect it with our backend.
One of the challenges we encountered during development was the time-consuming process of video compilation, which involved fetching the voices of each rapper, generating verses, and combining them into a video. To overcome this, we focused on core functionality and maintained open communication. Another challenge was the low-quality voices of some rappers, which affected the realism of our application. We addressed this by implementing longer-form-speech in our models, resulting in more realistic voices.
Accomplishment’s we’re proud of We integrated a modern UI into our front end, creating a modern and smooth experience for the user that allows them to create raps and easily interact with previously created raps. We made it possible to generate infinite rap battles between various rappers, allowing for endless topics, which allows users to customize their raps that are personalized to their interests. We utilized AI to develop a comical application that makes many people laugh out loud by implementing natural language processing models to produce verses that build off of responding to the opponent’s previous verses. We implemented video in our application, giving the rap battle a more engaging and realistic feel. Our team integrated blockchain wallets to create engaging, interactive, and decentralized voting procedures for users to express their ideas and opinions on which generated raps they enjoy.
What we learned We learned the most optimal way to interact between JavaScript and Python. This allowed us to utilize two major programming languages and transfer the data we needed between them effectively and quickly. We learned how to build an API that concatenated audio and a diverse set of technologies within artificial intelligence, such as text-to-speech and stable diffusion.
What’s next for SpitFire At SpitFire, our focus is on creating a platform that provides users with the most engaging and realistic rap battle experience possible. To achieve this, we are continuously exploring ways to enhance the existing features of our application. One such effort is the improvement of the voices of existing rappers. We understand the importance of realistic and authentic voices in bringing our rap battles to life. Our team is committed to ensuring that the voices of our rappers sound as close as possible to their real-life versions. By doing so, we believe that we can further enhance the user experience, making it even more enjoyable and engaging. Additionally, we are aware of the diverse array of celebrities and historical figures that exist and the importance of covering this range comprehensively. Our team is working tirelessly to add more rappers to our collection to achieve this. Our objective is to create a well-rounded and complete platform that caters to the interests and niches of as many users as possible. By adding more rappers, we aim to achieve this goal, providing users with endless possibilities for rap battles and enhancing the overall user experience of the SpitFire application. We are committed to expanding our user base and bringing more communities of people together through our application. We believe that a larger user base will allow us to create a more engaging and enjoyable experience for everyone. By supporting a large user base, we hope to foster a sense of community and bring people together from all walks of life. Our objective is to create a platform that is accessible and user-friendly for all, providing an enjoyable and inclusive experience for everyone. In addition, we plan to introduce a feature that allows users to become rappers themselves. This will create a more personalized and engaging experience among friends and family. By utilizing natural language processing models to mimic a user's voice and speech patterns, we aim to enhance the authenticity and fun of our application.