Hurry!CSIS Core Team Induction forms close on 30th August 2026 at 11:59 PM00h 00m 00s•NewCSIS Contributor Application Forms are now LIVE — Open to all batches!Apply as Contributor •
Hurry!CSIS Core Team Induction forms close on 30th August 2026 at 11:59 PM00h 00m 00s•NewCSIS Contributor Application Forms are now LIVE — Open to all batches!Apply as Contributor •
Hurry!CSIS Core Team Induction forms close on 30th August 2026 at 11:59 PM00h 00m 00s•NewCSIS Contributor Application Forms are now LIVE — Open to all batches!Apply as Contributor •
Hurry!CSIS Core Team Induction forms close on 30th August 2026 at 11:59 PM00h 00m 00s•NewCSIS Contributor Application Forms are now LIVE — Open to all batches!Apply as Contributor •
Back to Projects
Natural Language Processing

Ved

Visit Application
Project IDPROJ-005
StatusBeta Testing
Development Phase85% Complete
Lead DeveloperAnkur Kumar, Divyanshu Nath Tripathi, and Suyash Shukla

Context & Challenge

Finding precise guidelines within lengthy university regulation PDFs and documents is highly tedious for new students.

Core Purpose

Enable instant regulations assistance using a Retrieval-Augmented Generation model.

What was Built

A chat-based model supporting speech-to-text queries over indexed academic files.

Key Features

  • RAG model optimized for local web execution
  • Voice-call mode utilizing low-latency WebRTC streams
  • Collaborative notes canvas shared between peer accounts

Technologies Used

WebRTCTensorFlow.jsCanvas APICRDTSvelte

Want to contribute?

This project is built and maintained by CSIS members. Day scholars and freshers are welcome to contribute code, suggestions, or writeups.

Inquire via Contact