Machine output

lla provides stable JSON, NDJSON, and CSV listing modes for scripts and data pipelines. They preserve normal path, filtering, sorting, depth, and archive behavior; only rendering changes. JSON and NDJSON include plugin fields, while listing CSV uses the fixed schema below.

Formats

lla --json
lla --json --pretty
lla --ndjson
lla --csv

• --json streams one JSON array. --pretty only affects this mode.
• --ndjson emits one JSON object per line.
• --csv emits a header followed by data rows.
• The three output flags are mutually exclusive.

JSON and NDJSON fields

{
  "path": "src/main.rs",
  "name": "main.rs",
  "extension": "rs",
  "file_type": "file",
  "size_bytes": 1234,
  "modified": "2024-05-01T12:34:56Z",
  "created": null,
  "accessed": null,
  "mode_octal": "0644",
  "owner_user": "mohamed",
  "owner_group": "staff",
  "inode": 1234567,
  "hard_links": 1,
  "allocated_size_bytes": 4096,
  "xattrs": {"user.example": 12},
  "has_acl": false,
  "security_context": null,
  "mount_point": "/",
  "mount_source": "/dev/sda1",
  "filesystem": "ext4",
  "symlink_target": null,
  "is_hidden": false,
  "git_status": "M.",
  "plugin": {}
}

extension, timestamps other than modified, ownership, filesystem metadata, the symlink target, and Git status can be null. file_type is file, dir, symlink, or other. The plugin object contains enabled plugin fields.

CSV columns

CSV uses this fixed column order:

path,name,extension,file_type,size_bytes,modified,created,accessed,mode_octal,owner_user,owner_group,inode,hard_links,allocated_size_bytes,xattrs,has_acl,security_context,mount_point,mount_source,filesystem,symlink_target,is_hidden,git_status

Plugin actions have their own typed output modes. See Installing and managing plugins.

For selecting records before emission, see Filtering and search.

Search results

Content search uses a separate output contract:

lla --search "TODO" --json
lla --search "TODO" --ndjson
lla --search "TODO" --csv

For search, both --json and --ndjson pass through ripgrep's newline-delimited JSON event stream. It contains ripgrep event types such as begin, match, end, and summary; it is not the listing JSON array described above. --pretty does not transform this event stream.

Search CSV currently prints one line per collected match under these columns:

file,line,column,kind,text

This search-specific renderer does not fully escape commas or quotes in file names, so use the ripgrep JSON event stream when paths may contain those characters.

See Filtering and search for the search filters that are currently applied.