Skip to content

Aadit-17/Retrievus

Folders and files

NameName
Last commit message
Last commit date

Latest commit

Β 

History

12 Commits
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 
Β 

Repository files navigation

Retrievus - AI-Powered Knowledge Exploration Platform

A modern full-stack application that enables intelligent knowledge discovery using Pinecone vector database, Google Gemini AI, and role-based access control.

πŸš€ Features

  • πŸ” Intelligent Search: Multiple retrieval modes (Dense, Sparse, Hybrid)
  • 🎭 Role-Based Access: Different access levels (Intern, Employee, Manager, Executive)
  • πŸ€– AI-Powered: Pinecone vector database + Google Gemini response generation
  • πŸ“š Session History: Track searches during your session
  • 🎨 Modern UI: Beautiful React interface with smooth animations
  • ⚑ Real-time: FastAPI backend with live search capabilities

🎯 Usage

  1. Select your role (Intern, Employee, Manager, Executive)
  2. Choose retrieval mode:
    • Dense: Semantic similarity search
    • Sparse: Keyword-based BM25 search
    • Hybrid: Combined approach
  3. Enter your query and click "Generate Insights"
  4. View results with AI-generated summaries
  5. Check history panel for previous searches

πŸ” Search Modes Explained

Mode Description Best For
Dense Semantic embeddings Conceptual queries, synonyms
Sparse Keyword matching (BM25) Exact terms, technical queries
Hybrid Combined dense + sparse Best overall performance

🎭 Role-Based Access

  • Intern: Learning materials and basic information
  • Employee: Standard company policies and procedures
  • Manager: Team management and operational data
  • Executive: Strategic and high-level information

πŸ“š Session History

  • Automatically tracks all searches during your session
  • Click any history item to restore previous results
  • History clears when you refresh the page
  • Shows search metadata (time, mode, role, results count)

About

RAG Playground to demonstrate different retrieval types

Topics

Resources

Stars

Watchers

Forks