The-Vibe-Company/quivr ? reverse-engineered prompt

Build me a Python library that works like a second brain for my apps. I want to point it at a few files like PDFs, text files, and markdown, then ask natural language questions and get useful answers back from those files.

It should be easy to plug into an existing project, work with different LLMs like OpenAI, Anthropic, Mistral, or a local Ollama model, and let me customize the retrieval flow a bit, like changing how it searches and ranks chunks. Please include a simple way to create a brain from files and ask it a question in just a few lines of code, plus a small example I can run from the command line.

If you need to, look up current docs online for any model or API details, but keep the setup simple and focused on the core RAG experience.