FastAPI backend implementing a Retrieval-Augmented Generation (RAG) chatbot designed to answer questions based on local text content (page_content.txt
). Features automatic language detection and translation between English and Hindi.
- Framework: FastAPI (with streaming responses)
- RAG Core: LangChain
- LLM: Ollama (
phi3.5:3.8b-mini-instruct-q8_0
) - Embeddings: Ollama (
nomic-embed-text
) - Vector Store: ChromaDB
- Translation:
argostranslate
(Hindi <-> English),langdetect
- Knowledge Base: Sourced from
page_content.txt