Reference
crab clone
Clone a Crab repository from cloud storage.
Synopsis
crab clone [OPTIONS] <URL> [<DIRECTORY>]Description
crab clone clones a repository from a crab:// URL. The clone downloads
git objects (commits, trees, pointer blobs) but not the actual file content —
files appear as pointer blobs in the working tree until hydrated.
For conceptual background, see Cloning a Repository.
Arguments
| Argument | Required | Description |
|---|---|---|
URL | Yes | Repository URL (e.g. crab://my-bucket/my-repo) |
DIRECTORY | No | Target directory. Defaults to the repo name. |
Options
| Option | Default | Description |
|---|---|---|
--branch | default branch | Branch to check out |
--hydrate | false | Hydrate all files after cloning |
--hydrate-pattern | Glob patterns to hydrate after clone | |
--json | false | Emit structured JSON output |
Examples
Basic clone
crab clone crab://my-bucket/my-repoClone and hydrate everything
crab clone --hydrate crab://my-bucket/my-repoClone and hydrate specific files
crab clone --hydrate-pattern '*.safetensors' crab://my-bucket/my-repoClone a specific branch
crab clone --branch feature/v2 crab://my-bucket/my-repoRelated Commands
crab hydrate— hydrate files after cloning.crab init— create a new repository.crab fetch— fetch updates from remote.