Automate Audio Transcription with Zapier + Deepgram

Wouldn’t it be nice to be able to just drop an audio file into Google Drive and have a transcription of the file created automatically? With Zapier, a popular automation tool, it’s possible.

Before we dive in, why use Zapier for this in the first place? Because there are all kinds of things we could automate with Zapier that wouldn’t be possible otherwise. Here are some ideas:

• Transcribe voicemails and send them to your email or Slack

• Automate transcription for recorded meetings

• Transcribe customer support calls and create tickets

• Generate podcast show notes

The first step is to get familiar with Zapier and learn how to use their Webhooks integration so we can start building out all our great ideas. 

In this tutorial, we’ll start with a basic workflow -  drop a file in Google Drive, use Zapier Webhooks to send the file to Deepgram to be transcribed, and then add a text file of the transcription to a folder in Google Drive.

Creating an audio transcription zap

Let’s jump right into creating our zap. We’ll head to the Zapier dashboard. We’ll choose the classic view (versus the AI view) so we can tell Zapier which apps we want to use. To start, we’ll connect Google Drive to Webhooks by Zapier. This will give us a template to work with.

We can add the other steps to the workflow. We’ll click the plus sign to add another Webhooks by Zapier step (which will be a GET request), and lastly another Google Drive step (Create File from Text in Google Drive).

In Zapier terminology, the first step is the “Trigger”. This step is what will cause the entire workflow to run. We’re designing a zap that will kick off when we add an audio file to a specific Google Drive folder.

Step 1: Adding a file to a folder in Google Drive

Let’s configure the first step. We’ll click through the sections and complete the form to tell Zapier which Google Drive account to connect to and which folder to watch so that if an audio file is added to it, the zap gets triggered.

Now we can test the first step. This will run the step so that Zapier looks in the Google Drive folder to find audio files. It will show us the files it finds, and we can select one to serve as the example file. 

Important Note: If you select a file of a certain mimetype, such as .wav, Zapier will only trigger the zap when you add a file of that same mimetype. So decide at this point which audio file type you want to use.

Step 2: POST request with Zapier Webhooks

A webhook is a real-time communication method where one system automatically sends data to another system in response to a specific event, facilitating instant information exchange. 

Since Deepgram isn’t available yet in Zapier as an integration, we can use Zapier’s Webhooks to make a transcription request to Deepgram. In step two, we will tell Zapier to use a webhook to automatically send the audio file from Google Drive to Deepgram for transcription as an http POST request.

This step requires us to fill out a few important pieces of information in the form.

First, we’ll be sure to choose a POST request as the event and add Deepgram’s API url as the URL. For payload, we’ll leave that as form because that is the form the audio file will be in when it comes from Google Drive.

We can leave the Data input blank and stick with the default “no” for Wrap Request in Array. But we’ll have to make sure we select the correct input for the File. We must choose the Web Content Link from the Google Drive record. That will send the audio file in the appropriate format to Deepgram.

We don’t need to change the following two fields - Unflatten or Basic Auth. But we will need to add something to the Headers. This will be our authorization token, i.e. our Deepgram API key.

Now we are ready to test this step of the workflow. When we click “Test”, Zapier will attempt to send the audio file from step 1 to Deepgram to transcribe. When Deepgram sends it back, we should see a response like this:

Important Note: Zapier events will timeout after 30 seconds, so if we add an audio file in Google Drive, send it to Deepgram, and don’t get a response in 30 seconds, the zap will fail. We recommend splitting large audio files into smaller ones to avoid this timeout issue with Zapier.

Step 3: Create file from text in Google Drive

The final step is to put the transcription in a folder in Google Drive.

In this step, we’ll again need to connect our Google Drive to Zapier by adding our gmail account for that specific Google Drive account. Then for the action, we will add information such as the folder we want to put the transcription into, the name we’ll give the transcription file, and the file content. 

We’ll want to select the transcript property in the results from Deepgram that we received in the previous step. That will be the text that Zapier takes and creates a text file from:

We can test this step, and afterwards we can take a look in the transcriptions folder and we should see our transcription file!

Important Note: According to Zapier, if a file is directly added to the folder but was modified or created greater than four days earlier, the Zap won't trigger. This means only newer audio files added to Google Drive will trigger the zap to run. Read more at Zapier’s website: “Common Problems with Google Drive”.

Conclusion

Congratulations! Your Zapier-Deepgram integration now effortlessly transcribes audio files. The door to customization is wide open—explore and make it your own!

Sign up for Deepgram

Sign up for a Deepgram account and get $200 in Free Credit (up to 45,000 minutes), absolutely free. No credit card needed!

Learn more about Deepgram

We encourage you to explore Deepgram by checking out the following resources:

1. Deepgram API Playground 

2. Deepgram Documentation

3. Deepgram Starter Apps

More to Come

We are currently working on a Zapier Integration that will allow you to connect and facilitate the automation of tasks between different apps. An integration allows Zapier to interact with the APIs of various applications, enabling the creation of Zaps to automate workflows across those applications.