Automated Patent Abstract Summarization and Logging Workflow

This n8n workflow automates the process of summarizing patent abstracts, creating embeddings, and intelligently storing and retrieving data for analysis or review. The process starts with a webhook trigger, which receives patent abstracts for processing. The abstract is then split into smaller text chunks using a character-based text splitter. These chunks are converted into embeddings using OpenAI’s models and stored in a vector database powered by Supabase. When a new query or patent abstract comes in, the workflow queries the vector store to find similar previous abstracts. An AI agent then interprets the query or abstract, prepares a summarized response or relevant information, and logs the entire process into a Google Sheet for record-keeping. This workflow is ideal for patent offices, research labs, or any organization that manages large sets of patent data and needs automated summarization, similarity search, and logging capabilities.