SharePoint Connector - Technical Manual — Unique AI Documentation
SharePoint Connector - Technical Manual
The SharePoint Connector is a Node.js-based service that integrates Microsoft SharePoint Online with the Unique platform. It periodically scans configured SharePoint sites, downloads flagged documents, and ingests them into the Unique knowledge base.
Core Capabilities:
- Fetches content from Microsoft Graph API and SharePoint REST API
- Processes and forwards content to Unique AI's APIs
- Handles authentication through Azure AD (Microsoft Entra ID)
- Manages permission synchronization between SharePoint and Unique
- Operates on a configurable schedule (default: every 15 minutes)
Documentation
| Document | Description |
|---|---|
| Architecture | System components, hosting models, connectivity |
| Flows | Content sync, permission sync, file diff mechanism |
| Metadata | Ingested metadata fields for documents and SitePages |
| Permissions | Microsoft Graph and SharePoint REST API permissions |
| Security | Security updates, reports, and SBOM |
Key Concepts
Pull-Based Architecture
Unlike the previous Power Automate-based connector (v1) which used push-based principles, the SharePoint Connector v2 pulls information from SharePoint:
Benefits:
- No dependencies on Power Automate licensing
- Full control over sync timing and behavior
- Better visibility into sync operations
- Containerized deployment model
Selective Sync via Column Flag
Documents are not automatically synced. Users must explicitly flag documents by setting a custom column value:
File Diff Mechanism
The connector maintains state to detect changes:
- New files: Flagged files not previously synced
- Modified files: Files with changed content hash or metadata
- Deleted files: Previously synced files no longer flagged or deleted from SharePoint
See File Diff Flow for details.
Standard References
- Microsoft Graph API - Graph API overview
- SharePoint REST API - SharePoint REST service
- Microsoft Entra ID - Authentication and authorization