Opensource speech-to-text API for your app - [AI series]

Speech-to-text API

Speech-to-text can be used to transcript for instance a customer call, or another example, generate the subtitles of a video.

Let's see how simple it is to create an API with RYAX,
to transcript the video into a text file !

For that, we'll use Whisper, the open source library published by OpenAI,.

And ice on the cake, we'll use our chatGPT integration to generate the code

There's no code to write !

But aligned with our low-code philosophy, the code is still customizable, allowing any updates and an easy maintenance.

 

[AI series]

Read our tutorials on how to boost your application with opensource generative AI

>>> Let's start

First, let's create the action in Ryax, taking a video file in input and extracting the text

create ryax action from chatgpt

Click on "New run"

create ryax action from chatgpt 2

After few seconds, the workflow is fully executed , creating an action ready-to-use in others Ryax workflows

create ryax action from chatgpt 3
create ryax action from chatgpt 3
create ryax action from chatgpt 4

Now, let's use this new action just generated in another workflow, to transcript file and generate an API callable for this function.

This new workflow has 2 steps :

  • an HTTP POST API JSON step, to get the file to transcript
  • the new action created "Extract text from video file using Whisper"

Then click on "DEPLOY" to deploy it and make it available as API

It's that simple !

 

create ryax action from chatgpt 5
create ryax action from chatgpt 6

That's it !

The workflow is ready to use, through the UI, with a form, or through an API

If you're using API, we recommand to secure it by creating API key.

And thanks to RYAX offloading capacity, execution can be configured to use GPU if needed for any step !