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
Implemented pipeline
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.