Document Q&A with Voyage-Context-3 Embeddings & MongoDB

This n8n workflow implements a sophisticated Document Q&A system leveraging Voyage-Context-3 embeddings from VoyageAI and MongoDB Atlas as the vector store. It processes research papers (like Arxiv documents) by fetching, chunking, embedding, and storing them for efficient retrieval. The workflow is divided into two main parts: the ingestion pipeline that populates the vector database, and a RAG-based Q&A agent that retrieves relevant chunks to answer queries accurately.

Key benefits include superior embedding quality for context-heavy documents, seamless integration with MongoDB Atlas for scalable vector search, and a ready-to-use chatbot interface. It saves significant time on manual document analysis, enabling AI-powered insights from large PDFs or research papers without custom coding.

Ideal use cases: Academic research Q&A, enterprise knowledge bases, legal document review, or any scenario needing precise semantic search on lengthy texts. Customize by swapping the document URL in the 'Set Variables' node. Requires VoyageAI API key (with credits) and MongoDB setup; publish the workflow to activate the public chat UI for best performance.