Zen

(powered by Zengine)

Comment

💫 Inspiration

Everyone says "Just google it!", and nowadays, folks have started to say "just ChatGPT it!". Searching on google nowadays doesn't always give us the answer we're looking for, and gradually, more and more people are turning to chatGPT to answer their questions.

But not even ChatGPT can be the perfect search tool. It's a text-based LLM (Large Language Model) and can only return text information. Zen, provides the perfect balance between information and images, links and more for any user!

😮 What Zen does

Zen is an AI-focused search engine under a privacy-centric browser known as Zengine. It uses smart AI features such as Retrieval Augment Generation (RAG) models to create user-friendly, engaging content from ethically scraped websites.

🔨 How we built it

Through the power of Retrieval Augmented Generation, LLMs and Vector Databases, Zen Browser utilizes the power of RAGs and LLMs with a fine-tuned model that optimizes content from HTML to be more engaging. It does this in a privacy-friendly way and encourages developers to better their content.

😰 Challenges we ran into

  • Implementing Framer Motion! This was my first time using the library, and it required a fair bit of delving through documentation to find what I wanted. +Another issue that took fairly long to resolve was the prospect of the User's query being taken and sent from frontend to backend, where it would be fed into the AI
  • As the app scaled in the prototype, developing the front-end React view was becoming increasingly heavier.
  • Finally, the biggest issue was deploying the server to Github Pages.

😤 Accomplishments that we're proud of

We learned and explored various languages, libraries and frameworks over these past 36 hours. It has been an incredibly huge learning experience for us and we're proud to say that these acquired skills will be handy in the future!

🧠 What we learned

We learned…

  • How to scale our idea for the prototype
  • Large Language Models
  • Framer Motion
  • An approach to modularize solutions

💙 What's next for Zen

The current Zen search engine is only a fragment of the prototype. Much needs to be done! The crawlers should be updated to be more sophisticated and the policies regarding that and the scraper should be more rigid to follow ethical guidelines. The RAG should be refactored to allow for more flexibility to provide more personalized, local data to the user. The server needs to be fully fleshed out and refactored to be scaled and the frontend needs to be finished!

Built With

Share this project:

Updates