Cloud storage integrations
Cloud-storage connectors ingest files and folder hierarchies into the workspace graph. Each file becomes a _FileEntry node with a child _Blob (the binary), and the surrounding folder structure is preserved as _Folder nodes linked by _HasChild edges. Content extraction (text, tables, images) happens automatically in the background once a _Blob lands in the graph.
flowchart LR
Src[(Source bucket / drive)] -->|list| C[Connector]
C -->|metadata| FE[_FileEntry]
C -->|content| B[_Blob]
C -->|hierarchy| F[_Folder]
FE -->|HasContent| B
F -->|HasChild| FE
C -->|sharing / ACLs| AG[_AccessGroup]
AG -->|HasPermission| FE
Permission mapping is the differentiator
Sources that expose per-item sharing (Google Drive, OneDrive, SharePoint, Box, Dropbox) map those ACLs onto access groups; bulk-storage sources (S3, Azure, FTP) only map the integration-level configuration.