Read first

• This is all unofficial and only serves as an educational hack.
• Official SDK will probably be released in the future. You may want to wait.
• Only part of the API is implemented.
• Crypto is mostly ported from browser code and isn't optimal. Probably some of it could be done with openssl, ursa or cryptojs or the algorithms could at least be ported to use Buffer format, but this is no way a priority.
• Based on developer guide and site source.
• If you use it for something make sure you agree with Mega's Terms of Service.

Installation

npm install git://github.com/tonistiigi/mega.git

var mega = require('mega')

See examples directory for quick start.

Missing functionality

• No sharing features
• Server to client requests are not handled.
• Missing file management: move, mkdir, symlink etc.
• methods can only be called after ready event.

API

var storage = mega([email], [password], [readyCallback])

Create new connection instance to Mega. If you don't specify email/password then temporary account will be created. Once connection closes for temporary account you cannot access same account again so you need to save a link to file. Temporary accounts regularly get deleted.

After readyCallback() or ready event fires storage has following properties:

name - Account owner name

key - Account master key

rsaPrivateKey - RSA private Key

sid - Current session ID

files - Hash of File objects by node ID-s.

root - File object for Cloud Drive main directory

trash - File object for Rubbish bin

inbox - File object for Inbox

mounts - Array of all top level directories

storage.upload(options | name, [buffer], [cb])

fs.createReadStream('myfile.txt').pipe(storage.upload('myfile.txt'))

Upload a file to mega. You can pass in buffer data or just pipe data into it. Callback returns uploaded file object. If you don't specify callback you can listen for complete event to get the file handle.

Supported options:

name - File name required

attributes - Object of file attributes.

size - File size. Note that because Mega's API needs final data length before uploading can start, streaming only fully works if you specify the size of your data. Otherwise it needs to first buffer your data to determine the size.

target - Target directory file object or node ID. Defaults to storage.root.

storage.reload(cb)

Reloads files tree. No need to call this.

mega.file(url | opt)

var file = mega.file('https://mega....')

Returns file object based on download URL or options. Options can be downloadId and key.

File

Properties:

name - File name

attributes - Object of attributes

size - File size

key - File key(buffer)

timestamp - File creation time

nodeId - File ID

downloadId - Link ID to file. Only if created from link.

directory - Boolean if file is directory.

children - Array of files for directories.

file.download([cb])

Read file contents.

file.download().pipe(fs.createWriteStream('myfile.txt'))

file.download(function(err, data) {
  // data is buffer
})

file.link([noKey], cb)

Make download link for a file.

file.link(function(err, url) {
  // url: https://mega.co.nz/#!downloadId!key
})

mega.encrypt([key]) / mega.decrypt(key)

Lower level duplex streams. These could be used if you want to do network traffic and crypto on different time.

Takes in encrypted file data and outputs decrypted data and vice versa. Also does MAC verification / generation.

Note that if you specify key for encrypt() it needs to be 192bit. Other 64bit are for the MAC. You can later read the full key from the key property of the stream.