Skip to main content
Back to Search AI connectors list The Custom Content Connector enables Search AI to index content from applications that do not have a pre-built connector. It acts as a middleware between Search AI and a third-party application, retrieving content via REST APIs and converting it into an indexable format. A reference implementation of the connector service is available for download and serves as a template for integration. The current implementation supports basic authentication.

​
Setup

​
1. Download the Service

Obtain the skeleton custom connector service from the SearchAssist Toolkit on GitHub.

​
2. Configure the Service

  • Modify config.json to include the application configuration details and content fields.
  • Set the authorization password in the .env file in the root folder. Rename the existing .env.sample file to .env and add your auth value to the Authorization key.

​
3. Implement Changes (if needed)

Modify the service implementation as required. Refer to the readme file for details.

​
4. Host the Service

Host the configured service on your preferred platform.

​
Configure the Custom Connector in Search AI

  1. Go to Content > Connectors and select Custom Connector.
  2. Configure the following fields:
    FieldDescription
    NameUnique name for the connector
    EndpointURL and HTTP method of the hosted service. For the default implementation: GET https://<serverip>/getContent
    HeadersKey-value pairs sent as request headers
  3. Under Headers, add the authorization header required by the default service implementation:
    • Key: Authorization
    • Value: <Your-Auth-Key> (the value set in the service’s .env file)
    Note: The encoding format for a header can be set to base64 (value is encoded before sending) or none (value is sent as plain text). The default for newly added headers is base64.
  4. Click Connect to complete the setup.

​
Content Synchronization

  1. Go to the Configurations tab of the custom connector configuration page.
  2. Click Sync Now.
The service retrieves batches of 30 documents from the third-party application and sends them to Search AI. Search AI processes each batch and indexes the content. Synchronization is complete when all batches are processed.

​
Access Control

Search AI supports RACL on content ingested via the custom connector. The sys_racl field in the connector service config.json specifies which field in your source data holds permission information. The data fetched from that field is saved as a permission entity. For example, if access information is stored in a permissions field in your content, set the sys_racl field in the connector service config accordingly: Permission Field A permission entity is created for every value fetched from the specified field. Use the Permission Entity APIs to add users to each permission entity. Only users added to a permission entity can access the content associated with it. To enable RACL, go to the Permissions and Security tab and select Permission Aware. If Public Access is selected, all ingested content is available to all users.