Connect the systems that hold your company's knowledge, with permissions synced.
A source connects one system (or part of one) to one knowledge base. Every connector syncs two things: the content, and the access rights that go with it.
| Content | Indexed as |
|---|---|
| Documents (PDF, Office, Google Docs/Sheets/Slides, Markdown, text) | Text with structure (headings, tables); spreadsheets by sheet and range |
| Images inside documents, scanned PDFs | Text extracted by OCR; layout preserved where possible |
| Wiki pages and databases | Page text and properties; database rows as records |
| Tickets and conversations | Message threads with participants and timestamps |
| CRM/ERP records | Selected objects and fields as records |
| Database tables/views | Rows from the tables or views you select, read-only |
| Websites | Pages under the paths you include, crawled on schedule |
Files above 200 MB, binaries and archives are skipped and listed under Skipped items.
Each connector maps the source's permission model to principals (users and groups) stored with every document. Users are matched by email; groups are resolved at sync. See Permissions & access control.
Sources → Sync status shows last sync, items indexed, items skipped and errors. Failed items are retried automatically; persistent failures are listed with the reason.
Removing a source deletes its documents from the knowledge base within minutes and from backups as they rotate (35 days). Queries already logged keep their citations as text but the links stop resolving.
All connectors are listed under Integrations.