Create a transcription

Learn how to create an audio transcription with the Klangio API.

1. Model Selection

For demonstration purpose we will use the generalized universal model to process the audio file. To provide optimal transcription results, it is highly recommended to read the Model Selection for Transcription Jobs in Advanced Usage.

2. Decide on the Output Formats

The Klangio AI can generate Transcription jobs results in various formats, based on your needs. Results will be available for 14 days. Make sure to fetch and store them in your system or else there won't be any option to recover them, besides making a new request.

warning

Use as little output formats as possible. Multiple formats may inflict additional quota costs for a transcription job.

Available output formats:

• mxml (MusicXML)
• gp5 (Guitar Pro 5)
• midi
• midi_quant (quantized)
• pdf

3. Request via API

Create a transcription via Klangio API: (also see step 1 in Basic Job Workflow)

  import requests

  API_KEY = 'YOUR_KLANGIO_API_KEY'

  audio_file = open('{YOUR_FILE_NAME}', 'rb')

  query_parameters = {
    'model': 'universal', # required
    'title': 'Awesome recording', # optional
    'composer': 'John Doe', # optional
    'webhook_url': '{YOUR_WEBHOOK_ENDPOINT_URL}' # optional, but recommended
  }

  data_params = {
    'outputs': ['midi', 'pdf', 'gp5']
  }

  multipart_params = {
    'file': audio_file
  }

  resp = requests.post(
    'https://api.klang.io/transcription',
    headers={
      'kl-api-key': API_KEY
    },
    params=query_parameters,
    data=data_params,
    files=multipart_params
  )
  
  if resp.status_code == 200:
    print(resp.json())
    # {
    #  'job_id': 'SOME_JOB_ID', 
    #  'creation_date': '2023-01-01', 
    #  'deletion_date': '2023-02-01',
    #  'gen_xml': false,
    #  'gen_midi': true,
    #  'gen_midi_quant': false,
    #  'gen_gp5': true,
    #  'gen_pdf': true,
    #  'status_endpoint_url': 'https://api.klang.io/job/{SOME_JOB_ID}/status'
    # }

4. Check Job Status

See step 2 in Basic Job Workflow

5. Get the Output Data

Once the transcription job is in state COMPLETED or the webhook notification is received, you can fetch the output results from the job/{job_id}/{format} endpoint.

Here is an example python code snippet to fetch the xml of the transcription: (also see step 3 in Basic Job Workflow)

  import requests

  API_KEY = 'YOUR_KLANGIO_API_KEY'

  resp = requests.get(
    'https://api.klang.io/job/{YOUR_JOB_ID}/xml',
    headers={
      'kl-api-key': API_KEY
    })

  if resp.status_code == 200:
    with open('test.mxml', 'wb') as file:
      file.write(resp.content)