ti fs fork-layer
Forks a writable child timeline from a parent layer tip or one parent checkpoint. Forking pins the parent overlay without copying its file tree.
Syntax
ti fs fork-layer
--parent-layer-ref <string>
[--actor-id <string>]
[--checkpoint-id <string>]
[--dry-run]
[--file-system-id <string>]
[--fs-token <string>]
[--help]
[--layer-id <string>]
[--layer-name <string>]
[--version]
Options
--parent-layer-ref <string>: Parent layer ID, unique name, or supported tag reference. [required]--actor-id <string>: Actor ID identifying the child owner.--checkpoint-id <string>: Pin the child to this checkpoint of the parent. If omitted, pins the serialized parent tip.--dry-run: Validate the request without applying changes.--file-system-id <string>: Select the file system. You can also setTI_FS_FILE_SYSTEM_ID.--fs-token <string>: Set the file system user token. If omitted, usesTI_FS_TOKEN.--help: Display help information.--layer-id <string>: Stable child layer ID. If omitted, the service generates one.--layer-name <string>: Human-readable child layer name.--version: Display version information.
For options shared by all commands, see Global options.
Examples
Fork from the parent tip:
# Start an independent writable timeline at the parent's current serialized tip. ti fs fork-layer --file-system-id <file-system-id> --parent-layer-ref research-base --layer-name experiment-a --actor-id agent-aFork from a stable checkpoint:
# Continue from an earlier review boundary without changing the original timeline. ti fs fork-layer --file-system-id <file-system-id> --parent-layer-ref research-base --checkpoint-id seed --layer-name experiment-b --actor-id agent-b
Layer names remain visible after logical deletion and can become ambiguous. Automation should capture and use the returned layer ID.