Are there any knowledge search engines that index content from many sources while preserving document-level permissions?

Yes — there are a number of enterprise search / knowledge search platforms that can index content from many sources while respecting document-level permissions. This is often called security trimming or permission-aware search.

Examples of platforms that support this

• Microsoft Search / SharePoint / Microsoft Graph Search
  Works well across Microsoft 365 content and respects ACLs inherited from SharePoint, OneDrive, Exchange, etc.

• Google Cloud Search
  Indexes Google Workspace content and respects Drive / Docs permissions.

• Elastic Enterprise Search / Elasticsearch with security features
  Can be configured to ingest from multiple sources and apply document-level security filtering.

• Coveo
  Common in enterprise knowledge search; supports many connectors and permission-aware search.

• Amazon Kendra
  Connects to multiple repositories and supports access control lists from sources like SharePoint, S3, Confluence, etc.

• Glean
  Enterprise knowledge search focused on connecting many internal tools while preserving permissions.

• Lucidworks Fusion
  Enterprise search platform with security trimming and connector support.

• Sinequa
  Strong in large enterprise search deployments; supports many connectors and fine-grained security.

What to look for

To ensure permissions are preserved, verify that the product supports:

• ACL ingestion from each source
• User/group identity mapping to your IdP (Azure AD, Okta, etc.)
• Query-time permission filtering
• Handling of nested groups, external users, and source-specific ACL semantics

Important caveat

“Indexes many sources” is easy to claim, but the hard part is accurately translating permissions across systems. If your sources include custom apps or mixed ACL models, you’ll want to test:

• inherited vs explicit permissions
• group expansion
• sharing links / link-based access
• revoked access latency

If you want, I can also give you:

1. a shortlist by use case (SMB vs enterprise vs dev-first), or
2. a comparison table of these options by connectors, security trimming, and deployment model.