Chronicle Architecture

Workspace metadata, content storage, sync, and Fabric surfaces

Chronicle separates file metadata, content storage, filesystem presentation, and sync transport.

Component Layout

Chronicle
├── filesystem presentation
│   └── fuser-backed mount when FUSE is enabled
├── metadata store
│   └── SQLite file and iteration records
├── content store
│   └── BLAKE3-addressed content blocks
├── sync and replication
│   ├── local transport
│   └── WebSocket transport where enabled
└── optional Fabric server
    └── workspace/sync routes via --fabric-port

Metadata

Chronicle tracks directories, files, versions, deletion markers, permissions, timestamps, and content hashes. The metadata database is the queryable history of the workspace.

Content Addressing

File bytes are stored by content hash. This makes repeated content cheap, enables efficient sync, and lets different workspace states share identical blobs.

Sync

Sync moves metadata and content changes between locations. Consumers should treat sync as reconciliation, not as blind file copying.

Fabric Integration

When launched with a Fabric port, Chronicle can expose workspace and sync profile methods. Podium can attach Chronicle-backed workspaces to agent instances, while other services should respect Chronicle as the file-history authority.