Connectors
Available today
- SharePoint Online. Microsoft 365 SharePoint document libraries.
- Google Drive. Personal and shared drives.
- Confluence. Atlassian Confluence Cloud spaces and pages.
Dropbox, AWS S3, Azure Blob Storage, and Google Cloud Storage are on the roadmap.
How a connection works
Each connector uses OAuth 2.0. When you click Connect, you're redirected to the provider's sign-in screen, you authorise AnswerVault, and the provider issues a token that AnswerVault stores encrypted at rest (AES-256-GCM, per-tenant key derivation). AnswerVault never sees your password.
The token carries your access. AnswerVault can only see documents you can see. If your access changes — you leave a SharePoint site, a Confluence space is restricted — the connector picks that up on the next sync.
Important: Everything indexed in your tenant is queryable by every user in your tenant. The connection inherits your permissions, but the resulting knowledge base does not. Pick folders and spaces deliberately — don't connect anything you wouldn't share with everyone on the team.
OAuth scopes
AnswerVault requests least-privilege scopes — read-only where possible, no write access. The exact scope list is shown on the consent screen for each provider, and the full breakdown is on the security page.
SharePoint Online
Microsoft Graph delegated permissions for reading sites and files (Sites.Read.All, Files.Read.All). Tenant admin consent is typically required on the first connection in an organisation — the consent screen will say so. AnswerVault never requests Sites.ReadWrite.All or any write scope.
Google Drive
Google OAuth 2.0 with read-only Drive scopes (drive.readonly, drive.metadata.readonly). Works against personal drives and shared drives the connecting account has access to.
Confluence
Atlassian OAuth 2.0 with read scopes for spaces and pages (read:confluence-content.all, read:confluence-space.summary). Confluence Cloud only — Confluence Server / Data Center is not supported today.
Picking what to index
After OAuth, AnswerVault shows a browse view of what your account can see. Drill into the source and tick the folders, drives, or spaces you want indexed. Untick anything you don't.
Common patterns:
- Whole space. Useful for documentation-heavy Confluence spaces or a single SharePoint document library.
- Specific folders. Useful when a drive mixes shareable knowledge with personal or confidential content.
- Multiple sources, different scopes. Indexing HR's SharePoint site for policies plus Engineering's Confluence space for runbooks gives one unified knowledge base across both.
Click Save & sync when you're happy with the selection. AnswerVault starts a sync job in the background — see Syncs for what happens next.
Managing connections
The Connectors page in the app lists every connection in your tenant. From there you can:
- Add a new source on a different provider
- Re-open the picker on an existing connection to adjust what's indexed
- Disconnect a source — this removes the token and stops future syncs; already-indexed content can be cleared separately
Permissions and access control
Within a tenant, AnswerVault treats the indexed content as a single shared knowledge base. There is no per-document ACL replay at query time — if you don't want a document answerable, don't connect it.
Tenant isolation is hard: each tenant has its own encryption key, its own data residency region, and its own vector and graph stores. See the security page for the full architecture.