How can I convert binary data (Zoom transcript .vtt) into readable text in Make?

Questions
1

:bullseye: What is your goal?

I want to download a Zoom cloud recording transcript (.vtt file) in Make and convert it into readable UTF-8 text so I can extract the spoken content and use it in later steps (e.g., GPT processing, summaries, description generation, etc.).

:thinking: What is the problem & what have you tried?

:warning: Problem

When I download the transcript file via the Zoom module, the file arrives in Make as binary data.
The content is not directly readable — it looks like a hex/byte stream.

I need a way to decode this binary into actual text so that I can work with the VTT subtitle content.

:magnifying_glass_tilted_left: What I’ve tried

I tried using the module “Convert encoding of a text”, but since the input is binary, it does not decode the content.
The output looks identical to the input bundle and remains unreadable.

I also checked other Tools modules, but I haven’t found anything that takes binary input and returns text output.

:clipboard: Error messages or input/output bundles

Output Download-Modul Zoom: [
{
“fileName”: “Recording”,
“recording”: “IMTBuffer(2097213, binary, 90893d35b82f4ae4c2af9a4c99f893eda1bd3c4f): 7b2274696d656c696e65223a5b7b227473223a2230303a30303a30312e393630222c227573657273223a5b5d7d2c7b227473223a2230303a30303a35312e323130222c227573657273223a5b7b22757365726e616d65223a224d69636861656c20486162”
}
]

:camera_with_flash: Screenshots (scenario flow, module settings, errors)

Screenshot 2025-12-09 112852.png
Screenshot 2025-12-09 112852.png1031×775 54.7 KB

1 Like
3

Welcome to the Make community!

According to the official documentation, VTT files are in plaintext format.

Therefore, you can use the built-in function:

Hope this helps! If you are still having trouble, please provide more details.

— @samliew

4

Thank you Samliew. I tried the tostring function but the output was the same as the input. I didn´t get any text.

5

Are you very sure you’re downloading a VTT file??

It looks like Zoom’s recordings are in m4a (Audio format), try using a module like the OpenAI (ChatGPT, Sora, DALL-E, Whisper) “Generate a transcription” module —

Transcribes an audio to text.

For more information about the “Generate a transcription” module and the OpenAI (ChatGPT, Sora, DALL-E, Whisper) app, see the corresponding Integrations page and the Help Centre documentation.

Hope this helps! If you are still having trouble, please provide more details.

— @samliew

1 Like
6

Thanks for the suggestion!
Using the OpenAI Generate Transcription (Whisper or transcribe) module would normally be a great workaround — unfortunately in my case it doesn’t solve the problem.

The audio files coming from Zoom are much longer than the current input limits of the Make → OpenAI Whisper module.

  • Whisper in Make has an audio length limit of 1,400 seconds (≈ 23 minutes)

  • My Zoom recordings are 40–90 minutes, so the module rejects the file

  • Splitting the files is not an option in this automation, because Zoom delivers them as a single M4A

So while Whisper would work for shorter audio, it can’t process these full-length recordings inside Make due to the time limit.

7

Ok so the file you are trying to get is the actual audio recording of the meeting and not the transcript of it?

8

Welcome to the Make community!

Try using the CloudConvert “Convert a File” module to compress/reduce the filesize.

For more information about the “Convert a File” module and the CloudConvert app, see the corresponding Integrations page and the Help Centre documentation.

If you are still having trouble, please provide more details.

— @samliew

1 Like
9

I need the transcript because I use it in OpenAI to generate a YouTube headline and a YouTube description. That’s why having the transcript is so important for my workflow.

Creating the transcript directly from the audio file is not an option, because the OpenAI module has a strict audio-length limitation, so the full recordings can’t be processed that way.

10

Yeah, and what I mean is your screenshot seems to be showing the audio file and not a transcript. And due to Make’s built in timeout limits and data transfer limits it would be better to try compressing the file first, like Sam suggested. Or try an external service to do the transfer.

2 Likes