AI Legal & Financial Document Analyst

Grounded AI answers from your legal and financial PDFs.

DocuMind combines secure PDF upload, text extraction, intelligent chunking, ChromaDB vector search, Groq-powered answer generation, and persistent document chat in a Django production architecture.

RAG Workspace

Document Intelligence

Version 1.0

Implemented pipeline

PDF extractionReady
Chunk embeddingsReady
Persistent chatReady

Source-grounded

Answers are built from retrieved document chunks.

User-scoped

Documents and chats are isolated per authenticated user.

Features

Built capabilities

DocuMind Version 1.0 includes the complete single-document RAG workflow from upload to persistent chat.

Secure Authentication

Login-protected dashboard, documents, retrieval, and chat.

PDF Upload

Validated PDF uploads with user-owned document records.

Automatic Text Extraction

Page counting and text extraction using the document processing service.

Intelligent Chunking

Text is split into retrieval-friendly chunks with overlap.

Semantic Search

Embeddings and ChromaDB power document-scoped retrieval.

AI Chat

Groq-generated answers are saved in persistent document chat sessions.

Source Citations

Assistant messages include source previews from retrieved chunks.

Conversation History

Saved sessions can be reopened from the history page.

User Isolation

Users can only access their own documents and chats.

System

Implemented RAG pipeline

The backend is organized as a service-driven pipeline with replaceable retrieval, vector store, and LLM layers.

PDF Upload

Validation

Text Extraction

Chunk Generation

Embeddings

ChromaDB

Semantic Retrieval

Prompt Builder

Groq LLM

Grounded Answer

About

Production-oriented Django RAG architecture.

DocuMind is a Version 1.0 AI portfolio product built with Django, Tailwind CSS, ChromaDB, Sentence Transformers, and Groq. The project emphasizes clean service boundaries, authenticated document ownership, grounded answer generation, and persistent chat sessions.