Binary Input Loader
The Binary Input Loader is a document-loading node in n8n that takes binary file data (PDFs, images, Word documents, spreadsheets) and converts it into a text format that AI models and vector databases can process. It sits in n8n’s AI document-loading chain and handles the first step of any retrieval-augmented generation (RAG) pipeline: getting unstructured files into a usable text format.
The core problem it solves is simple. Businesses have knowledge locked inside files: policy documents, contracts, technical manuals, invoices. To make that knowledge searchable by an AI agent or chatbot, those files first need to be parsed into text and split into chunks. The Binary Input Loader takes a binary file from an earlier node (an upload, an email attachment, a file read from cloud storage) and extracts the text content so downstream nodes can embed it into a vector store.
We use this node in AI agent development projects where clients want their AI assistant to answer questions from internal documents. It works particularly well in combination with n8n’s Text Splitter and Vector Store nodes. Feed it a PDF from Google Drive, and within a few nodes you have searchable, AI-queryable content without writing any extraction code.