Amazon's Cloud Drive offers unlimited cloud storage but no good way to interact with your data (upload, download, find, etc). The web app is lacking and the desktop app is sub-par. So here's a command-line interface and SDK for interacting with Cloud Drive as if it were a filesystem.
npm install -g clouddrive
Before using the CLI, the config values for the application will need to be set. Use the config command to view and set the available options.
$ clouddrive config
auth.email =
auth.id =
auth.secret =
cli.colors = true
cli.ignoreFiles = ^(\.DS_Store|[Tt]humbs.db)$
cli.progressBars = true
cli.progressInterval = 250
cli.timestamp = false
database.driver = sqlite
database.host = 127.0.0.1
database.database = clouddrive
database.username = root
database.password =
display.date = modified
display.showPending = true
display.showTrash = true
download.checkMd5 = true
json.pretty = false
log.file =
log.level = info
sync.chunkSize =
sync.maxNodes =
upload.duplicates = false
upload.checkMd5 = false
upload.numRetries = 1
$ clouddrive config auth.email me@example.com
email saved
You will need to set the email for the Amazon account you wish to use with the CLI. The first run of the application will require you to run clouddrive init to authorize the CLI with your Amazon account. This will open a browser and take you to Amazon for authorization. After authorization, your access token will be printed in the browser. Simply copy and paste this back into the terminal.
Optionally, if you'd like to use your own Amazon Cloud Drive credentials, set the auth.client-id and auth.client-secret options using the config command.
$ clouddrive init
Initializing...
Initial authorization is required
https://www.amazon.com/ap/oa?client_id=...
? url:
Naviage to the URL displayed to to authorize the app with your Cloud Drive account using your credentials. This will redirect your browser to a new URL: paste that URL back into the prompt.
The first time you run the CLI (after initialization), you will need to (and periodically after the initial sync) run the sync command to pull down any Cloud Drive changes to the local cache. This local cache is required for the CLI to work and speeds up reading information when 'browsing' Cloud Drive using the CLI. This also makes many commands available for offline use.
$ clouddrive sync
The CLI makes interacting with Cloud Drive feel like using a remote filesystem with commands such as ls, du, mkdir, etc.
Usage:
clouddrive command [flags] [options] [arguments]
Commands:
about Print app-specific information
cat Print files to STDOUT
clearcache Clear the local cache
config Read, write, and reset config values
delete-everything Remove all files and folders related to the CLI
download Download remote file or folder to specified local path
du Display the disk usage (recursively) for the specified node
exists Check if a file or folder exists remotely
find Search for nodes by name
info Show Cloud Drive account info
init Initialize and authorize with Amazon Cloud Drive
link Link a file to exist under another directory
ls List all remote nodes belonging to a specified node
metadata Retrieve metadata of a node by its path
mkdir Create a remote directory path (recursively)
mv Move a remote node to a new directory
pending List the nodes that have a status of "PENDING"
quota Show Cloud Drive account quota
rename Rename a remote node
resolve Return the remote path of a node by its ID
restore Restore a remote node from the trash
rm Move a remote Node to the trash
share Generate a temporary, pre-authenticated download link
sync Sync the local cache with Amazon Cloud Drive
trash List the nodes that have a status of "TRASH"
tree Print directory tree of the given node
update Update a node's metadata
unlink Unlink a node from a parent node
upload Upload local file(s) or folder(s) to remote directory
usage Show Cloud Drive account usage
Global Flags:
-h, --help Show help [boolean]
-v, --verbose Output verbosity: 1 for normal (-v), 2 for more verbose (-vv),
and 3 for debug (-vvv) [count]
-q, --quiet Suppress all output [boolean]
-V, --version Show version number [boolean]
The config command is used for reading, writing, and resetting config values for the CLI. The following options are available:
auth.email: The email to use with the CLIauth.id: Custom Amazon API credentials if you would like to use your ownauth.secret: Custom Amazon API credentials if you would like to use your owncli.colors: ANSI color outputcli.progressBars: Display or suppress progress barsdatabase.driver: Database type to use for the local cache (sqlite,mysql, ormongo)database.host: Host/IP the database exists on (if not SQLite)database.database: Database to use (MySQL)database.username: Database username for authenticationdatabase.password: Database password for authenticationdisplay.date: Display eithermodifiedorcreateddate when listing nodesdisplay.showPending: Toggle displaying ofPENDINGnodes withlscommanddisplay.showTrash: Toggle display ofTRASHnodes with thelscommanddownload.checkMd5: Perform or suppress MD5 check when downloading filesjson.pretty: Whether to format JSON output or notupload.duplicates: Allow duplicate files to be uploaded to Cloud Driveupload.retryAttempt: Number of attempts to upload a file
The ls command allows you to view the contents of a folder. If you don't provide a remote path argument, it will display the contents at the root directory. The output provides detailed information for each item including the remote ID, its modified date (or created if you change the config), its status (AVAILABLE, TRASH), its type (FILE or FOLDER), its size, and its name.
You can also provide a node's ID instead of path for the ls argument by using the -i flag.
$ clouddrive ls
1234564789 Nov 8 15:20 AVAILABLE FOLDER 0B Documents
0123456829 Mar 30 8:35 AVAILABLE FOLDER 0B Pictures
8723457923 Aug 23 15:39 TRASH FILE 0B test.txt
The du command will output the total size used by the given file or folder, recursively. Again, if no argument is given, it will calculate the entire used space of your entire Cloud Drive. Passing the -i flag will calculate the size of the node by its ID instead of its path. It will also output the total files and folders contained in the path.
$ clouddrive du
174.77MB
3 files, 1 folders
The upload command lets you upload files and folders (recursively) to Amazon. Simply pass an arbitrary number of local paths (globbing is supported) and the last argument must be the remote folder to upload the files to. If you want to upload to the top-level directory, simply pass in / as the last parameter.
$ clouddrive upload ./test/* /
In addition to uploading files, the download command allows you to retrieve files you've uploaded to Amazon. The first parameter is the file or folder (recursively) you want to download. The second (optional) parameter is the location and/or filename to save the file as. If no path is given, the remote node is downloaded to the current working directory with the same name as it exists remotely.
$ clouddrive download /test/ .
You can optionally choose to encrypt your files upon upload to Amazon Drive. Passing the --encrypt flag with the upload command will encrypt both the file's name and contents before uploading to Amazon using the crypto options in the config. If no password is specified in the config, or you manually pass the --password flag, you will be prompted for the encryption password.
The encrypted file contents, by default, are a binary format generated by the crypto library in NodeJS. You can choose to run the file's contents through a base64 encoding after encryption to armor the contents by passing the --armor flag on upload. This setting is also sticky based on changing the crypto.armor config option.
Once files are encrypted, their nodes are tagged with the 'enc' label in Amazon Drive. This makes the app friendly across other commands such as ls, tree, etc. Using any of these commands with the --decrypt flag will decrypt the file's names to easily be read back out while listing and navigating through directories.
Passing the --decrypt flag with the download command will also decrypt the files on download so they exist locally, unencrypted. Note: you will need to use the same amoring flag or config setting when downloading files as you did uploading.