what is praello-ai-knowledge-assistant-for-businesses fr?

Praello is a full platform for building an AI knowledge assistant for a business. It lets a company upload its documents, such as PDFs, Word files, and Markdown, and then ask questions about that knowledge through an AI agent that searches and reasons over the uploaded material, a technique commonly called retrieval-augmented generation, or RAG for short. The system is built with a layered structure. A web frontend built with Next.js talks to a backend API built with FastAPI in Python. Behind that API sit several layers: one that handles the actual business logic and use cases, one that defines the core data and rules of the system, and one that manages the AI work itself, including an agent built with LangGraph, the RAG search pipeline, and safety checks called guardrails. Underneath all of this, the system stores data in PostgreSQL, uses Redis for caching, and uses Qdrant to store the embeddings, which are numeric representations of text used for search. The platform includes user accounts with role-based permissions, so a company can control who is allowed to upload documents, query the knowledge base, or manage the system as an admin. Roles range from a basic viewer up to a super admin with full access. There are API endpoints for uploading and managing documents, querying the knowledge base, running conversations with the AI agent, and checking system health and usage metrics. Everything runs through Docker Compose for easy setup, with Nginx acting as the entry point in front of the API and frontend. The project includes automated tests and continuous integration through GitHub Actions, along with code quality tools like Ruff and Black. A getting started guide walks through cloning the project, adding an OpenAI API key, and starting everything with a single Docker command. The project is released under the MIT license.