Connect External Knowledge via OpenAPI

Overview

Before you start

• Access permission: you need knowledge-ingestion management permission and must complete application authentication according to the current OpenAPI documentation.
• Version: use the current Knowledge Base 2.0 interface.
• Knowledge identity: clearly define the external knowledge object and its unique identifier.
• Knowledge attributes: prepare title, content, attachments, category, permission, update time and validity period.

Ingestion flow

1. Identify the knowledge object in the external system, including its unique identifier, title, content, attachments, category, permission, update time and validity period.
2. Complete application authentication and interface-call configuration according to the current OpenAPI documentation.
3. Map external fields to knowledge entry fields and determine the target space and folder.
4. Design rules for create, update, delete, failure retry and duplicate handling.
5. Validate with a small amount of test data: content, permission, status, attachments and AI usage results.
6. Enable formal synchronization and monitor failures and exceptions through logs.

Boundaries to keep

• Do not hard-code interface addresses, credentials or keys in documentation; follow the current OpenAPI documentation for specific parameters.
• The visibility scope of external knowledge must not exceed the authorization scope of the source system.
• The external unique identifier should be stable to avoid repeated duplicate knowledge from the same source.
• A successful import does not mean the knowledge is usable; it still must satisfy the published status, validity period and permission conditions.

Expected results and verification

1. Reopen the page with the current account.
2. Confirm that the test data enters the knowledge base as expected and keeps the permission boundary.

FAQ