Projects/RAG Platform

AI RAG Middleware Platform

A production-ready, enterprise-grade RAG (Retrieval-Augmented Generation) control plane that sits between applications and Large Language Models.

Production Readyv0.1.0

Technology Stack

FastAPI
Framework
Python 3.11+
Language
PostgreSQL 17
Database
Qdrant
Vector DB
Redis
Cache/Queue
MinIO
Storage
Angular 19
Frontend

Key Features

File Management

  • Support for PDF, DOCX, TXT, MD, HTML
  • S3-compatible object storage
  • Async background processing
  • Real-time status tracking

Embedding Pipelines

  • Configurable chunking and overlap
  • Multi-model support
  • Hybrid search (dense + sparse)
  • Pipeline versioning

Chat & Search

  • RAG-powered conversational AI
  • Server-sent events streaming
  • Session management
  • Hybrid retrieval with RRF

Multi-Tenancy

  • Organization-level isolation
  • User roles (Owner, Admin, Developer, Viewer)
  • API key management
  • Comprehensive access logging

Use Cases

Enterprise knowledge bases

Customer support automation

Legal and compliance document search

Research paper analysis

Interactive product documentation

System Architecture

🌐 API Gateway

FastAPI-based REST API with authentication, rate limiting, and comprehensive access logging

⚙️ Processing Layer

Async worker service for file processing, embedding generation, and vector indexing

💾 Data Layer

PostgreSQL for metadata, Qdrant for vectors, MinIO for files, Redis for queue

Ready to Get Started?

Deploy your own RAG platform or learn more about integration