Simple RAG Consumption with Databricks Mosaic AI > Introduction to Simple RAG Consumption with Databricks Mosaic AI recipe > Prerequisites
  

Prerequisites

To use the Simple RAG Consumption with Databricks Mosaic AI recipe, the following prerequisites must be met:
  1. 1Existing database table
    2. You must have a database table available on which the vector index search will be performed.
  2. 2Enable datafeed
    3. Enable the datafeed for this database table to allow continuous data ingestion.
  3. 3Create a vector endpoint
    4. In Databricks, create a vector endpoint. A Databricks vector endpoint is a scalable, serverless API hosted on Databricks that enables real-time serving of vector embeddings generated by machine learning models.
    The following image shows the option to create a vector endpoint:
    The image shows an option to create a vector endpoint.
  4. 4Create a vector search index
    5. Under the created vector endpoint, create a vector search index for the database table.
    The following image shows the vector search index option:
    The image shows the vector search index option.
    The following image shows a sample creation of a vector search index:
    The image shows the Create vector search index page.
After creation, the vector endpoint will appear on the Vector Search page as shown in the following image:
The image shows the list of vector endpoints on the Vector Search page.