Kuika's Search action allows you to search database or file collections based on a specific query. The Search action takes different parameters depending on the type of search and returns results according to the criteria specified by the user.
The Search action can be used in three different ways:
Vector Search - A search type that returns results closest to the given query.
Exact Vector Search - Returns vectors containing an exact query.
RAG Search (Retrieval-Augmented Generation) - It is the most suitable action for question and answer and works integrated with OpenAI.
In this tutorial, you will learn how to use different types of searches.
The Search action is supported on web and mobile applications.
Search Action Usage Details
Limit value can be specified for Vector and Exact Vector Search.
Collection Name can be optionally added so that only the specified collection is searched.
Vector Search
Returns the closest results to the given query.
The results are sorted from closest to farthest.
If no limit value is given, it returns up to 5 results by default.
Exact Vector Search
Returns results that match the entered query exactly.
For example, if the word “Izmir” appears in full, only these results are returned.
RAG Search (Retrieval-Augmented Generation)
It is the most appropriate action for question and answer purposes.
Thanks to the integration with OpenAI, it creates the most appropriate answer by combining the given query with the documents in the collection.
Add Search Action
Log in to the Kuika platform.
Open the project you will work on from the Apps screen.
While in the UI Design module, open the Properties panel on the right side.
Select Initial Actions → Searching → Search from the +ADD ACTION drop-down menu.
In the settings panel that opens, configure the following fields:
Example Configurations
Vector Search
Objective: To find sentences containing the word “Izmir” or similar sentences.
Search Type: Vector Search
Query: izmir
Limit: 5 (Up to 5 results will be returned.)
Exact Vector Search
Objective: To find complete sentences containing the word “Izmir”.
Search Type: Exact Vector Search
Query: Izmir
Limit: 5 (Up to 5 results will be returned.)
RAG Search
Objective: Getting an answer to the specified question using OpenAI with documentation from the collection.
Search Type: RAG Search
Query: What is the old name of Izmir?
(Assume that there is an information file about Izmir in the system.It will return the most appropriate answer according to this query).
After completing these steps, the Search action will be successfully added and you can offer your users a fast and smart search experience.
