Automate ADAS Event Annotation with n8n and Langchain

Overview

This workflow is designed to streamline the annotation process for ADAS events. It leverages the power of n8n's automation capabilities combined with Langchain's advanced NLP features. The process begins with a Webhook receiving incoming ADAS event data. This data is then processed by a Text Splitter node to prepare it for embedding. Cohere Embeddings generate vector representations of the text, which are then inserted into a Supabase vector database using the Vector Store Insert node. For retrieval and analysis, a Query node fetches relevant data from Supabase, which is then processed by a Langchain Agent. The Agent utilizes tools and memory to understand and act upon the ADAS event information. Finally, the results are logged to a Google Sheet for further review and action. This workflow solves the problem of manual and time-consuming ADAS event annotation by providing an automated, scalable, and data-driven solution.