Inspiration
Team ByteForce aims to create a content management software for a non profit organization called NMSTA which uses musical therapy to help individuals with disabilities. The services the non profit conducts includes therapy, music lessons, and professional development. A challenge that NMSTA faced was organizing their Google Drive as they simply had so many files that it seemed impossible to track every single one of them and distribute them to different types of users. As a result, they were looking for a software that could organize their chaotic google Drive and allow them to send their files to any user with ease.
What it does
Content Management
- Automative Tagging
- Content Classification ## How we built it How we built it The Backend/APIS/databases includes. Django SQLite HuggingFace Rest Framework
The AI includes. Meta-LLama-3 OpenAI Whisper OpenAI CLIP
The Front End includes. Next.js React Tailwind shadCN TypeScript
Challenges we ran into
Some challenges that we struggled with was when we creating our frontend with specifically shadCN components. The documentation was not ideal and it made it extremely difficult to customize our software. Next, Tailwind's theming would prove to be an issue to us because we had to manually adjust the settings which provided to give much difficulty to us. Finally, making a custom datatable with Next.js was extremely hard because the document was once again horrible and amped up the difficulty of making the software.
Accomplishments that we're proud of
One of the accomplishemnts that we are proud of is that we managed to integrate three unique AI models including a LLM, CNN, and Transformer model all witin our software integrating perfectly with each other. Another implementation we are proud of is we the fact that we designed a fully functional back end and front end that allows improves effiency by enhancing file searching, indexing, and content management which all in turn leads to a streamlined user experience.
What we learned
Some tasks that we learend was how to implement different AI models using HuggingFace within the backend of a software. Furthermore, we gained a deeper understanding of how we can use Tailwind, Next.js, shadCN to make good looking software. Finally, we learned how to use Django to make a backend functional and render Rest APIs to serve the front end.
What's next for Librarian CMS
Eventually, we plan to integrate Google Drive into our software to manage files within Drive as well as normal files. Furthermore, we plan on implementing a chatbot that allows the user to chat with the files on the CMS for instant insights on widespread data.
Log in or sign up for Devpost to join the conversation.