Comment on page

Translate and Transcribe Audio

You can use Whisper in Takomo:

Step 1: Setting Up Your Project in Takomo

Begin by creating a new project in Takomo. Upon launching the Builder, you will notice a template section. While there's a ready-to-use template for translating speech to English, starting from scratch will provide a thorough understanding of the system. So, create a new project.

Step 2: Building Your Translation and Transcription Pipeline

After creating your project, you will see an empty pipeline. This pipeline comprises an input section (where you add the audio to be translated and transcribed), the output section (where the translated and transcribed text appears), and the pipeline itself that connects the two.
  1. 1.
    Add an audio input node to the input section and upload your audio file.
  2. 2.
    To set up the output, simply add a text output node.

Step 3: Incorporating the Whisper Node

The key element of your pipeline is a tool in the audio and speech section known as 'Whisper'. This node is responsible for translating and transcribing your audio. You can choose between a regular Whisper node and a Whisper Advanced node, with the latter offering more options but taking longer to run. For this guide, we will use the regular Whisper node.
Once you've added the Whisper node to your pipeline, connect it to both the input and output sections. This configuration allows the Whisper node to process the audio from the input and send the translated and transcribed text to the output.

Step 4: Translating the Audio to English

One of the valuable features of Whisper is its dual ability to translate and transcribe. If your audio is in a foreign language, you can use the 'Translate to English' option. This feature can be particularly useful for automating the translation of foreign-language customer support requests into English.

Step 5: Testing Your Pipeline

After you've set up your pipeline, it's time to test it. Simply click 'Run Pipeline' and wait for the process to finish. Upon completion, the translated and transcribed text will appear in the output section. You can view this output as a text file or connect it to other software via Takomo's API.

Step 6: Integrating the Output with Your API

If you wish to integrate the output with your software, access the API Pane in Takomo. This section displays all API settings, including the models used in your pipeline, the inputs and outputs, and the price per API call. You can generate the API from this section. Note that once the API is generated, the pipeline locks to prevent editing while the API is being deployed.
To integrate the API with your software, navigate to the API documentation located in the right corner. This will provide you with all the necessary information, including post requests, get requests, and responses.