I’m building a small web app that feels as effortless as ChatGPT yet is tailored for product look-ups and file retrieval. A visitor should be able to type something like “show me lightweight Bluetooth headphones under $100” and immediately receive concise product details. The same conversational window must also let the user say “open the installation guide for model X” and get a clickable link to the exact text, PDF, or image file stored on the server. Core behaviour • Natural-language understanding powered by an LLM (OpenAI API or similar) combined with vector search (FAISS, Pinecone, etc.) so relevant products and files surface quickly. • A chat box with multimedia support—bold or italic text, inline images, and link previews. • All conversations, along with user IDs and timestamps, are saved in PostgreSQL so anyone can pick up where they left off. Preferred stack Python with FastAPI is my default choice, but Flask or Django are fine if you already have scaffolding in place. Front-end can be pure HTML/JS or React; whichever route you choose, responsiveness and simplicity are key. Deliverables I will review before releasing the single milestone: 1. Web-based chat UI served over HTTPS that accepts natural language, shows formatted responses, and embeds clickable file links. 2. Python backend that receives chat messages, hits the LLM, performs vector lookups, and returns structured answers. 3. PostgreSQL schema and migration scripts for storing users, chat history, products, and file metadata. 4. Setup guide (README) plus launch script or Dockerfile so I can reproduce the environment locally. Acceptance criteria • Queries for any existing product return at least name, price, key attributes, and a confidence score. • “Show me the PDF/manual/image” style commands yield a valid URL the browser can open directly. • Conversation history is loaded correctly after page refresh or new login. • End-to-end tests pass in my environment without manual tweaks. Payment is released only after the above items are fully delivered; partial drops won’t be considered. If this matches your expertise, let’s get started!