Azure Cognitive Services and Power Automate

I needed to transcribe some audio files to text. So, I took this as an opportunity to learn a bit about Azure Cognitive Services and figure out how to use Power Automate to complete this task.

You can download the Flow from here - https://link.freefall365.com/cognitiveservicesfiles.

Or from the Power Automate cookbook - https://powerusers.microsoft.com/t5/Power-Automate-Cookbook/Convert-Audio-Files-to-Text-Using-Azure-Cognitive-Services-Batch/td-p/441482

I documented the method I used to learn and summarised the steps taken in the following four short videos.

I found the whole process thoroughly enjoyable and felt empowered once I achieved my goal.

From a no-developer, I felt way out of my comfort zone at times, but by the end, I felt empowered and eager to learn more about cognitive services and Power Automate.

Now that I can use the HTTP Request action in Power Automate with the REST API in Azure Cognitive Services, what other Azure services can I use? The possibilities are HUGE!

Part 1: The Prerequisites.

In this part, I look at

• What Service I Used How did I decide what steps to use in Power Automate?

• What key info I took from the MS documentation

• How I got a Cognitive services subscription key

• How I got my audio file into Azure Blob storage and how I got the URL of that file

• Where to find the correct REST API endpoint to use and where to see an example of the request I should send to it.

You can download the Flow from here - https://link.freefall365.com/cognitiveservicesfiles. This is video one in a four video series where I demonstrate how to use Power Automate to transcribe audio files using Azure Cognitive Services Batch Transcription feature and my thought process through learning about the steps involved. Part 1: The Prerequisites.

Part 2: Creating the Transcription Request

This part covers

• How I used the HTTP Request action in Power Automate to create a Batch processing request in Azure Speech to Text service

• What URI to use How to authenticate with Azure cognitive services What I used for the body of the request

• What valuable information we get back from the request and how to analyse the output of the action in Power Automate

You can download the Flow from here - https://link.freefall365.com/cognitiveservicesfiles. This is video two of in a four video series where I demonstrate how to use Power Automate to transcribe audio files using Azure Cognitive Services Batch Transcription feature and my thought process through learning about the steps involved.

Part 3: Retrieving the Transcription request, checking the status and results

In this part I summarise

• How I parsed the JSON response that we got back from the POST request.

• How to retrieve the transcription

• How I used the Do Until action in Power Automate to poll the transcription and check it's status periodically I did some housekeeping and deleted the Transcription from Azure Cognitive Services.

• And finally how I analysed the output of a completed transcription request.

You can download the Flow from here - https://link.freefall365.com/cognitiveservicesfiles. This is video three in a four video series where I demonstrate how to use Power Automate to transcribe audio files using Azure Cognitive Services Batch Transcription feature and my thought process through learning about the steps involved.

Part 4: The Results

And finally, in this part, I demonstrate the following -

• How I parsed the JSON responses from getting a completed transcription

• How I retrieved the transcription result from Azure Cognitive services

• Some of the issues I obstacles I overcame while trying to analyse the output from our transcription result and get it into a format that I can save to a text file

• How I saved the text file to OneDrive

You can download the Flow from here - https://link.freefall365.com/cognitiveservicesfiles. This is video four in a four video series where I demonstrate how to use Power Automate to transcribe audio files using Azure Cognitive Services Batch Transcription feature and my thought process through learning about the steps involved.