Lemme Learn More (LLM)

logo
mobile responsive_2
mobile responsive_3
mobile responsive_1
web_3
web_2
web_1

Inspiration

In a world full of so much info but limited time as we are so busy and occupied all the time, we wanted to create a tool that helps students make the most of their learning—quickly and effectively. We imagined a platform that could turn complex material into digestible, engaging content tailored for a fast-paced generation.

What it does

Lemme Learn More (LLM) transforms study materials into bite-sized TikTok-style trendy and attractive videos or reels, flashcards, and podcasts. Whether it's preparing for exams or trying to stay informed on the go, LLM breaks down information into formats that match how today's students consume content. If you're an avid listener of podcasts during commute to work - this is the best platform for you.

How we built it

We built LLM using a combination of AI-powered tools like OpenAI for summaries, Google TTS for podcasts, and pypdf2 to pull data from PDFs. The backend runs on Flask, while the frontend is React.js, making the platform both interactive and scalable. Also used fetch.ai ai agents to deploy on test net - blockchain.

Challenges we ran into

Due to a highly-limited time, we ran into a deployment challenge where we faced some difficulty setting up on Heroku cloud service platform. It was an internal level issue wherein we were supposed to change config files - I personally spent 5 hours on that - my team spent some time as well. Could not figure it out by the time hackathon ended : we decided not to deploy today.

In brainrot generator module, audio timing could not match with captions. This is something for future scope.

One of the other biggest challenges was integrating sponsor Fetch.ai agentverse AI Agents which we did locally and proud of it!

Accomplishments that we're proud of

Biggest accomplishment would be that we were able to run and integrate all of our 3 modules and a working front-end too!!

What we learned

We learned that we cannot know everything - and cannot fix every bug in a limited time frame. It is okay to fail and it is more than okay to accept it and move on - work on the next thing in the project.

What's next for Lemme Learn More (LLM)

Coming next:

realistic podcast with next gen TTS technology
shorts/reels videos adjusted to the trends of today
Mobile app if MVP flies well!

Built With

fetch.ai
flask
llms
moviepy
openai
python
react
text-to-speech

Submitted to

YHack Fall 2024
- Winner FetchAI

Created by

Backend: I worked on the brain rot module. I used moviepy library and open AI chat completion API to extract information from data.

Aditya (Addy) Sahu
Passionate Software Engineer with 3 years of experience in backend development - looking to break into the US Market. Help Appreciated! :)
I built Flashcard microservice using Flask.
It takes an input document containing knowledge and generates flashcards that helps student to memorize faster and interact with the flashcard.

Alisha Tak
I designed and developed a podcast microservice using Flask and integrated it with OpenAI's GPT4 and Google's Text-To-Speech. The service handled complex audio file uploads, processing, and automated tasks such as summarization, and content generation. Implemented robust file handling pipelines, efficient communication with external APIs.

Aakash Kardam
Doctoral Student | UMass Dartmouth | Computer & Information Science
I worked on the frontend part using ReactJS and built the whole UI of the project. I integrated all the backend APIs built on flask with the frontend code. Also, I made the website mobile responsive and helped my team mates to solve critical bugs related to API integration.

Pulkit Kalra