Anyone played around with OpenAI whisper in n8n?

[ramzi_t]

Describe the problem/error/question

What is the error message (if any)?

Please share your workflow

1. I’m trying to get it to transcribe an audio I get from the whatsapp API but get this error : I tried every possible option…

I might not be getting the right output from Whatsapp?

2. [10:59 AM]

Share the output returned by the last node

Information on your n8n setup

• n8n version:
• Database (default: SQLite):
• n8n EXECUTIONS_PROCESS setting (default: own, main):
• Running n8n via (Docker, npm, n8n cloud, desktop app):
• Operating system:

[Mr.caribou]

Hey @ramzi_t I’m very interested in seeing the answer to this as well. I just started building a Whatsapp business cloud workflow today, but stuck at the webhook → respond to webhook step. The workflow executes successfully when I click to verify the webhook within Facebook API, but then when I try to send test messages either within Facebook API or directly from my phone, nothing. The workflow just keeps trying to listen for triggers.

[Jon]

Hey @ramzi_t,

Welcome to the community

I have taken a quick look at your screenshot and the OpenAI documentation and the first thing I noticed was you are sending an OGG file which doesn’t look to be supported.

The audio file object (not file name) to transcribe, in one of these formats: mp3, mp4, mpeg, mpga, m4a, wav, or webm.

Could this be the issue? The rest of the call looks good to me.

[ramzi_t]

Could be. I’ll try to convert it.

But I did the same automation and it works with ogg without conversion.

[Jon]

Hey @ramzi_t,

It may have worked but the API docs don’t mention it so it would be worth sticking to the documented formats. It could also be that the data is being sent wrong but might as well start with the basics and go from there.

[marvinwu]

we were able to make it work:

the work flow is send a mp3 to whisper for transcribe, and it worked without much problems

here is the node setting: