Skip to content

README.md

Latest commit

 

History

History
47 lines (34 loc) · 1.75 KB

README.md

File metadata and controls

47 lines (34 loc) · 1.75 KB

InstaLite

InstaLite is an AI-powered social media platform that enables personalized photo sharing, actor face matching, and intelligent chatbot conversations. Built as a full-stack application, it brings together real-time user interactions, cloud storage, and retrieval-augmented AI capabilities.

Overview

InstaLite allows users to:

  • Sign up, upload a profile picture, and receive celebrity face matches using facial recognition.
  • Share and browse posts through a real-time feed.
  • Interact with an AI-powered chatbot that references user-uploaded documents to provide accurate, context-aware responses.

Features

  • Secure authentication using JWT and bcrypt
  • Face matching with 46,000+ actor embeddings using vector search
  • Cloud-based image upload and retrieval via AWS S3
  • AI chatbot powered by LangChain and document-aware RAG pipeline
  • Real-time post feed and user interaction features
  • Hashtag-based search and user recommendation system

Tech Stack

Frontend:
React.js, TailwindCSS, Axios

Backend:
Node.js, Express.js, MySQL, ChromaDB, AWS SDK

AI and Infrastructure:
TensorFlow.js for face embeddings
LangChain for RAG pipeline
Kafka for streaming

Face Matching Workflow

  1. User uploads a profile photo.
  2. The image is uploaded to AWS S3.
  3. A facial embedding is generated using a face recognition model.
  4. Embedding is compared to a database of actor embeddings stored in ChromaDB.
  5. Top matches are returned and displayed on the user profile.

Chatbot Architecture (RAG)

  1. User enters in a question/statement/any form of inquiry.
  2. Text is chunked, embedded, and stored in ChromaDB.
  3. At query time, LangChain retrieves relevant chunks.
  4. A language model generates a context-aware answer based on retrieved content.