Transcribing with a different Acoustic or Language Model

    You can use a different/custom Acoustic Model and/or Language Model which better matches your audio data in order to get higher accuracy.

    Specifying the Acoustic Model

    Below is a table of all the publicly available Acoustic Models at this time. You can learn more about Acoustic Models, and other Acoustic Models we plan to launch, on the Custom Models page.

    Model Use Case Parameter
    Default      Best for all types of data and accents. English only. "assemblyai_default"                                                             
    Australian-English                                 Tailored for Australian-accented English. You can expect a 2-3% improvement compared to the default model on Australian data. "assemblyai_en_au"
    Custom Tailored for your data. You can expect a large improvement compared to the default model on your data. Contact us for more info: support@assemblyai.com

    To swap the Acoustic Model, just specify the "acoustic_model" param name in the POST to /v2/transcript.

    curl --request POST \
      --url https://api.assemblyai.com/v2/transcript \
      --header 'authorization: YOUR-API-TOKEN' \
      --header 'content-type: application/json' \
      --data '
        {
            "audio_url": "https://s3-us-west-2.amazonaws.com/blog.assemblyai.com/audio/8-7-2018-post/7510.mp3",
            "acoustic_model": "assemblyai_en_au"
        }'

    Specifying the Language Model

    You can use the API to train a custom Language Model which will include any custom vocabulary, like product names and person names, unique to your application. You can also boost accuracy for common phrases and speaking patterns. For example, speaking patterns in a conversation versus a lecture/presentation are very different, and a Language Model can capture this difference.

    To use your custom Language Model in a request, just set the "language_model" param to the name of the Language Model you created. For example:

    curl --request POST \
      --url https://api.assemblyai.com/v2/transcript \
      --header 'authorization: YOUR-API-TOKEN' \
      --header 'content-type: application/json' \
      --data '
        {
            "audio_url": "https://s3-us-west-2.amazonaws.com/blog.assemblyai.com/audio/8-7-2018-post/7510.mp3",
            "language_model": "my_custom_lm"
        }'
    Heads up! When using a custom Language Model, processing time can slow down by as much as 25%.

    You can mix and match Acoustic Models and Language Models to get the best setup for your application and data.