← Back to portfolio

AI / Web App

AI Document Analyzer

AI Document Analyzer
AI-powered document analysis platform based on Retrieval-Augmented Generation (RAG). The system allows users to upload documents such as PDFs, DOCX, TXT, and CSV files, process them through an AI pipeline, and interact with their content through natural language queries.

The application performs a full document processing workflow including text extraction, cleaning, semantic chunking, metadata generation, and embedding creation. The processed data is indexed in a vector database to enable fast semantic search and contextual question answering.

Users can chat with their documents, generate summaries, and extract relevant information grounded directly in the source content. The system also supports conversational memory for follow-up questions and streaming responses for real-time interaction.

Architected a modular backend using FastAPI and Python, implementing a complete RAG pipeline with vector search, reranking capabilities, and multiple LLM providers including Groq, OpenAI, Gemini, and local models through Ollama. Built the frontend with Next.js, React, and Tailwind CSS, creating an intuitive interface for document upload and AI-assisted exploration.