Morpheus

Intelligent document reasoning system with semantic search and source citations. Upload private documents, ask questions in natural language, and receive accurate answers with citations.

Project Overview

Morpheus is a Retrieval-Augmented Generation (RAG) system featuring a Matrix-themed interface. Users can upload private documents and ask questions in natural language, receiving accurate answers with source citations from their uploaded content.

The system uses session-based isolation with fresh Pinecone vector namespaces per user, ensuring complete data privacy. When sessions end, all data is automatically deleted with no permanent storage.

Built with a focus on cost efficiency, Morpheus offers a token-based pricing model compared to traditional subscription alternatives, making it accessible for researchers and developers.

Technical Stack

Frontend: Next.js, TypeScript, Tailwind CSS
Backend: Python, FastAPI
AI/ML: Claude (Anthropic), OpenAI embeddings, LangChain
Database: Pinecone vector database
Deployment: Vercel (frontend), API backend

Key Features

Document Upload

Upload and process multiple document formats for intelligent querying

Natural Language Q&A

Ask questions in plain English and get accurate, cited responses

RAG Modes

Multiple retrieval modes including semantic search and hybrid approaches

Session Isolation

Complete data privacy with session-based namespaces and automatic cleanup

Vector Search

Semantic search using Pinecone for accurate document retrieval

Fast Performance

Optimized chunking and embedding for quick responses

How It Works

1. Document Processing

Documents are chunked and converted to vector embeddings using OpenAI's embedding models

2. Vector Storage

Embeddings are stored in Pinecone under session-specific namespaces for isolation

3. Semantic Retrieval

User queries are embedded and matched against stored vectors for relevant context

4. Response Generation

Claude generates context-informed responses with source citations

5. Session Cleanup

All data is automatically deleted when sessions end, preventing persistence

6. Privacy First

No permanent storage ensures complete data privacy and security

Project Screenshots

Implementation Highlights

Frontend Architecture

  • Next.js with TypeScript for type-safe development
  • Matrix-themed UI with Tailwind CSS
  • Real-time chat interface for document querying
  • Responsive design for all devices

Backend Processing

  • FastAPI for high-performance async operations
  • LangChain for RAG pipeline orchestration
  • Document chunking with overlap for context preservation
  • RESTful API design for clean integration

Privacy & Security

  • Session-based isolation with unique namespaces
  • Automatic data deletion on session end
  • No permanent storage of user documents
  • Secure API key management

RAG Implementation

  • Pinecone vector database for semantic search
  • OpenAI embeddings for document vectorization
  • Claude for context-aware response generation
  • Multiple RAG modes for different use cases