You can easily integrate Search actions into the applications you develop on the Kuika platform. These actions enable vector search on resources added to the system, allowing the most relevant documents to be identified with high accuracy in response to user queries.
In addition, these documents can be sent to large language models such as OpenAI using the RAG (Retrieval-Augmented Generation) method, enabling the generation of contextual and direct responses to user queries.
Searching actions:
Works with files, folders, or sources in base64 format.
Converts these sources into vector format by dividing them into pieces of specific sizes.
Each piece is securely stored and linked to metadata.
You can delete sources at any time or keep them up to date by synchronizing them at specified intervals.
This allows you to make data access and search operations faster, more efficient, and more user-friendly.
Search Actions Available in Kuika
Save Source
Allows you to upload files from Google Drive, AWS S3, or manual providers and save them to the database.
You can search the saved files.
Get All Source List
Lists all sources associated with a specific Collection ID.
If no collection name is specified, sources in the Default Collection are listed.
Delete Source
Deletes sources with the specified Collection Id from the system.
Allows you to clean up unnecessary or old sources.
Search
Performs a search in the database or file collections based on a specific query.
Can be used with different search types such as Vector Search, Exact Vector Search, and RAG Search.
List Collections
Lists all collections in the Qdrant system, enabling them to be used as values in projects.
List Sources
Lists sources in the Qdrant system, enabling them to be used as values in projects; optional parameters can be used to list sources belonging to specific collections.
You can integrate these actions into your application to implement advanced search and data management operations on the Kuika platform.