Getting Started with Search & RAG

How to Start Using Search & RAG Easily

Once your data is ingested and properly processed (see Data Ingestion), you can begin using the Search & RAG feature to search for specific data or ask questions about it.

The simplest way to explore Nuclia's search and generative answer capabilities is through the search page in your Nuclia dashboard. This page offers all available options, and by clicking the "Get code" button, you can obtain the corresponding API call or Python code to use in your application.

API Endpoints for Search & RAG

/find Endpoint

To search for data, use the /find endpoint. It combines semantic and keyword search to identify the most relevant data for your query. For more details, see Tune the Search Strategy.

Example:

/find?query=How to ride a bike

This will return the most relevant data for the query "How to ride a bike."

/ask Endpoint

To ask questions about your data, use the /ask endpoint. It internally calls the /find endpoint to gather appropriate results and passes them to the chosen LLM. Thus, the /ask endpoint accepts the same parameters as the /find endpoint, with additional parameters for answer generation. For more details, see Define the RAG Strategy.

Example:

/ask?query=How to ride a bike

This will return a generative answer to the question "How to ride a bike" based on the results returned by the /find call.

Deploying Search & RAG in Your Application

The easiest way to integrate Nuclia's search and generative answer capabilities into your application is by configuring a widget from your Nuclia dashboard and pasting the corresponding HTML snippet into your application.

Alternatively, you can implement a solution yourself using the API directly or our SDKs for Python or JavaScript/TypeScript.