Document Intelligence — Ask Your Files

Overview

Document Intelligence is a working proof-of-value for grounded document question-answering. Authenticated demo users upload a PDF or text-based file and ask questions conversationally; answers are kept strictly grounded in the uploaded document. The architecture is deliberately LLM-agnostic — the same flow runs against a local model served by Ollama (for example Llama 3) or any cloud LLM. This particular deployment is wired to Google Gemini, but the model is a swappable component, not a lock-in. Files are processed in-memory and expire automatically.

Problem Solved

Teams waste time hunting through long PDFs, policies, contracts, and reports. This demo shows how an LLM, properly grounded, can answer precise questions from a single document while refusing to fabricate answers that are not present — and how the model choice (local or cloud) can be swapped without changing the application.

Capabilities

• PDF and text-based file upload (txt, md, csv, json, log, xml, html, yaml)
• Natural-language question answering
• Strict document grounding (refuses unsupported answers)
• Conversational multi-question thread
• Pluggable LLM backend — Ollama / Llama (local) or cloud LLM
• Gated access with password + email OTP
• In-memory processing with automatic expiry
• Rate limiting and size limits to protect the service