Skip to main content

Bulk File Upload (SFTP)

To connect to AudioShake's SFTP server, users can download an SFTP Desktop Client like Cyberduck or Winscp. The server can be accessed using the user's AudioShake clientId and token. A config.json file with specified output format, licenseId, and names is required. A callbackUrl can be optionally provided for status updates.

Quick Start

  1. Download an SFTP Desktop Client: Cyberduck, Winscp.
  2. To connect to our SFTP server at sftp://sftp.audioshake.ai, enter your AudioShake clientId and token into the user and password fields respectively.
  3. Once authorized, please provide a config.json file with the following information. Ensure that in the config.json file, you specify the desired output format, a licenseId, and names (which represent the output stem types). Optionally, you could provide a callbackUrl to get up to date info on the status of your request.
  4. To kick off jobs, please add files in the input/ folder.
  5. Output files will be in the output/
    1. Folder structure of the output folder would be song_name/model_name.format
    2. Desktop Client: Cyberduck, Winscp
    3. SSH: You can use the sftp command in your terminal or command line interface. The format would be sftp yourclientId@sftp.audioshake.ai.

Config.json

The config.json file is a configuration file that is essential for using the SFTP server. It contains key information required for processing your request. The licenseId field is where you provide your license ID. The format field specifies the desired output format, which can be either wav, mp3, or flac. The optional callbackUrl field can be used to provide a URL where status updates on your request can be sent. Finally, the names field should contain a list of the output stem types you want, such as vocals, drums, etc.

Set the residual flag to true to return a separate residual file. This is an optional property.

{
"licenseId": "<LICENSE_ID>",
"format": "<wav, mp3, flac>",
"callbackUrl": "<OPTIONAL CALLBACK URL>",
"residual": false,
"names": [
"vocals",
"drums"
]
}
ElementDescription
licenseIdYour unique license id
formatThe desired output file format
callbackUrlA webhook site where multiple updates about the job status will be sent
residualWhether or not to also generate & output a residual, the source track discluding the given model
namesAn array of data models to use. A unique job will be created for every model
  • Please follow along here to see our models
  • Please follow along here to see our supported formats
  • If you're conducting a transcription or alignment job, set the format to json and use either transcription or alignment in the names array.
    • Do not include other model types in the names array. When the desired output format is 'json', you can only use either transcription or alignment.