Project June is actually an exploratory project aimed at understanding AI-integrated web development, and for a more focused view: how LLM wrappers are constructed. From a user's POV, the project is just an agent that can help one brainstorm a particular document. So, essentially, this is a TYPICAL project, nothing special or innovative but just a good way to get started with these uncharted concepts of the new era.
If the 'why' is about the name 'Project June': The project was a summer initiative for me in June, which is why it is called 'June'.
If the 'why' is about the motivation that drove me: The project was fuelled by the void that existed in my mind when the term 'AI' emerged. So, to fill that void, Project June was incepted.
Tho, there is no solid reason for anyone to install it, still if one insists on taking a look at how things are wired inside, one can follow these commands:
git clone https://github.com/suu-b/Project-June.git
cd Project-June
cd ./client
// add .env for client using ./client/.env.example
npm run dev //to run client
cd ./server
// add .env for server using ./server/.env.example
npm run dev //to run development server
The tech stack is MERN with Langchain for AI integration, HuggingFace for vector store and embeddings, and Gemini for LLM and NLP. Moreover, as an add-on, I've implemented an external search facility for the LLM so that it can respond to real-time queries, for that, serpapi is used. For parsing of markdown responses and HTML components, ReactHtmlParser and marked are used. And for great UI components, shadcn was a great help.
There is a compare documents feature pending which will be implemented shortly. Apart from this, I've a few broad ideas in my mind which would take a considerable time to motivate me to plan and eventually implement them. After their implementation, this project wouldn't remain a TYPICAL one but a thoughtful, refined, and well-implemented TYPICAL one.