Companies QA Chatbot with RAG
Project type: AI chatbot
The project
A RAG-based question-answering chatbot for companies: users upload PDF files and ask questions about their content in natural language.
A question-answering chatbot system for companies, built on the Retrieval-Augmented Generation (RAG) architecture. Users upload PDF files and then ask the chatbot questions, which it answers from the content of those documents. The system uses LangChain, a FAISS vector database, the OpenAI API for embeddings and text generation, FastAPI for the backend and React with Ant Design for the frontend, all containerized with Docker and Docker Compose.
Key features
- PDF files upload
- Natural-language Q&A over document content
- Embeddable website chat widget
- Containerized deployment with Docker and Docker Compose
Artificial intelligence
RAG architecture
Documents are split, tokenized and vectorized with OpenAI embeddings, then indexed in FAISS to retrieve the relevant passages for each question.
LLM orchestration
LangChain and LangGraph orchestrate retrieval and answer generation, with dedicated prompt engineering.
Answer monitoring
Langfuse traces conversations and model calls to monitor answer quality.
Screenshots
