Introduction

These set of articles will go over the suite of endpoints available to prompt the serviced models as well as query the vector store. Layar API offers the ability to do RAG retrieval and model prompting separately or together.

/layar/gpt/generate

The generate endpoint allows you to prompt any model being serviced on your environment. It gives you the flexibility to decouple RAG from prompting or to provide RAG queries at the time of prompting.

For more information on this endpoint: Understanding the Generate Endpoint

/layar/gpt/retrieval

The retrieval endpoint allows you to both generate vectors from text as well as search for vectors using a query. This endpoint has one destination /generateVectors.

/generateVectors

Allows for more control over embedding. Any serviced embedding model can be utilized to create vectors our of raw text.

For more information on this endpoint: Understanding the Vector Generation Endpoint