How nabhpatodi10/research-ai works
Project type and purpose
• This is a full-stack AI research platform (“Research AI”) combining a React/Tailwind/Firebase frontend with a FastAPI backend that orchestrates LangChain/LangGraph agents and chains against multiple LLM providers.
• Users sign up/login via Firebase, submit research questions, and get AI-powered analyses.
File structure
• Top-level directories:
• backend/ – FastAPI application, chain/agent modules, graph builder, scrapers, DB layer, Dockerfile.
• frontend/ – React/Vite application, Tailwind styles, Firebase auth/storage, components by feature (Auth, Chat, Feedback, etc.).
• Config files: .env.example in both; backend/requirements.txt; frontend/package.json, vite.config.js, vercel.json.
• Images subfolders hold architecture and UI mockups.
README
• Confirms architecture: FastAPI + LangChain/LangGraph on backend, React + Firebase on frontend.
• Mentions support for multiple LLMs (OpenAI, Groq, Gemini) and a “Research Graph Architecture.”
Detail without overcomplicating
• Every module you see: custom_search, scrape, graph, chains, agents, database.
• How they interconnect in the orchestration layer.
• Feedback loop: user feedback flows back into the backend for model improvement or logging.