Over here, we will explain in more details of the settings.


Please provide the endpoint for Typesense on the first field. If you are running a self-hosted Typesense, you should refer to their official documentation and enter the endpoint accordingly.

As we are using instantsearchJS, the Search API key should be the API key with only search permission and not the master API key or admin key. The Master API key allows unauthorized developers to write to Typesense. (Refer to Typesense Documentation)

Collection Name (id)

This is the name of your collection that you store in Typesense server. By default, if you are using our GPAI portal plugin, the collection id is ai-use-case.

Refer to the Typesense Collection Documentation for more information.

The Query By

This field is optional. If leaves blank, by default, we will query by ‘post_content’. This will allow us to display on the frontend when the search form loads.

The FastAPI details

Search API: This API endpoint is used during the search queries in the frontend. It uses the GET method to perform the vector conversion. (Refer to FastAPI Search Method)

Upload API: This is used during the indexing/uploading of documents (AI Use Cases) from WordPress to Typesense. It uses the POST method to perform the conversion. (Refer to FastAPI POST method)

Logging of Search Queries

This is an option if you want to log queries. The logging is based on a custom API endpoint built into GPAI Typesense Plugin. The logging message refers to the message shown in the frontend below the search box.

Hits Per Page

The hits per page is to determine how many items to show per pagination. Refer to official InstantSearch documentation.


This is referring to the vector dimension. This is used during schema creation during the initial indexing.

Threshold (new)

Stringent – Displays only the most precise results; this setting could yield no results depending on the search query used.
Balanced (Recommended) – Displays relevant results; some less relevant results may also be shown.
All Results – Displays all solutions available in the database, sorted by their relevance to the search query.

Caution: Semantic search engine currently works better for English queries and solutions.