How to Auto-Transcribe Audio Files with Whisper in Make.com

Voice notes and call recordings pile up untranscribed. This scenario watches a Google Drive folder, sends any new audio file to OpenAI Whisper, and saves the transcript as a text file beside it.

• A Make.com account
• A Google Drive folder for audio uploads
• An OpenAI API key (Whisper access is included)
• Audio files in a supported format such as mp3, m4a, or wav

Step 1: Watch the upload folder

Add the Google Drive Watch Files in a Folder module. Select the folder where you will drop recordings. Choose to watch by Created date so each new file triggers exactly once.

Step 2: Download the file binary

Whisper needs the actual file data, not just a link. Add a Google Drive Download a File module right after the watcher and map the file ID from the trigger. This passes the binary along the flow.

Step 3: Add the Whisper module

Add OpenAI Create a Transcription (Whisper). For the file input, map the data from the Download module. Leave the model as whisper-1. If you know the language, set it to skip auto detection and speed things up.

Step 4: Save the transcript

Add a Google Drive Create a File module. Set the file name to the original name with a .txt suffix, set the content to the Whisper text output, and point it at a transcripts folder.

File name: {{1.name}}.txt
Content:   {{3.text}}
Folder:    /Transcripts
Convert a document: No

Step 5: Test it

Upload a short clip to the watched folder and run the scenario once. Confirm a matching .txt file appears in your transcripts folder with readable text inside.

Result: any recording you drop in the folder becomes a searchable text transcript within a minute, with zero manual typing.