Convert Tour PDFs to Vector DB with Google Drive, LangChain & OpenAI

This n8n workflow streamlines the ingestion of tour package PDFs from a Google Drive folder into a Pinecone vector database. It begins with searching and downloading PDFs, then loops through each file to extract text using a data loader (e.g., pdf-loader or Unstructured). The text is cleaned, recursively split into manageable chunks, embedded with OpenAI models, and upserted into Pinecone for semantic search capabilities.

Key integrations include Google Drive for file handling, LangChain for document processing and chaining, OpenAI for high-quality embeddings, and Pinecone for scalable vector storage. Setup requires API keys for these services and a designated Drive folder with tour brochures containing details on activities, itineraries, and pricing.

Benefits include massive time savings in building RAG (Retrieval-Augmented Generation) systems for travel apps, enabling features like intelligent tour recommendations or chatbots. Ideal use cases: travel agencies creating AI search over brochure libraries, tourism platforms enhancing discovery, or developers prototyping vector-based Q&A systems without manual data prep.