LearnLink AI

Talk with your documents. LearnLink. Every document uploaded to LearnLink will be give a custom chatbot trained on the documents. Chatbot supports 95+ languages for easy accesability around the world.

Comment

Inspiration

Our inspiration for this project was driven by the frustration that big AI models will give wrong answers when it does not have the information. Our goal was to provide an easy way to customize and train chatbots on your own documents. We saw a huge potential use case in the Education, Finance, Healthcare, and Academia fields as all of these fields need really specific information for AI to be useful. We also had a goal of including the entire world to our project because we want people who do not speak English to

What it does

Allows for anyone to talk with documents. Just upload the document and it will be given a custom chatbot instance trained on the text and data from it so you can easily talk with and ask questions. You can also switch the language to get the information in the one that you want. Available in 95+ languages. This can be easily used on Textbooks (Education), Financial Reports, Medical Reports, Research Papers and many other documents to easily extract information that you need.

How we built it

This project has 2 parts, the frontend (website) and backend (REST API). The frontend is built with Next.js, React.js, TypeScript, Tailwind CSS, Node.js, and Material UI which follows Google Universal Design System to provide a beautiful UI with fast performance. The backend was built with Python, Flask, Flask API, OpenAI, and Langchain. The frontend is hosted on Vercel and the backend is hosted on Google Cloud Platform.

Challenges we ran into

Some challenges we ran into was the reading of the documents to train the chatbot. The process was to chunk them and store them in a vector database, but some issues came with the chunking which took a while to fix. Another issue was with the languages feature. Sometimes the language translation would not be correct, so we had to fix that.

Accomplishments that we're proud of

Produced a full application with a friendly UI and fully functioning backend. This is the first time in a while that an application had an everyday use case for us, and we plan on using this throughout my education. We are also was proud of the multilple language implementation as it allowed multiple people, including our parents, to be able to use the application. We were also proud of creating dynamic routes (shareable links) as we have never used them before.

What we learned

We learned how to create a fullstack project from scratch as we only ever had experience in frontend or backend separately. We also learned that a friendly UI is very important as everyone has to be able to use the application, not just us. We also learned that an idea is never impossible to execute as long as we are working together as a team and motivated.

What's next for LearnLink AI

Collaborate with textbook manufacturers to create premade chat bots, so that students dont have to upload their own textbook pdf files. Create voice-based interactions, in order to make the platform more accessible to people with disabilities

Built With

Share this project:

Updates