How to transcribe audio with OpenRouter in n8n (without using the official OpenAI/GPT node)?

Questions
1

Hi folks,

I’m currently using n8n to automate audio transcription I have an audio file coming in and the OpenAI “Transcribe a Recording” node works great for me. Now I’d like to replace OpenAI with OpenRouter to handle the transcription, but I don’t want to use the standard GPT or OpenAI node**.

So I’m stuck—here’s what I’d love to know:

  1. Does OpenRouter support Whisper-style audio transcription at all?
  2. If yes, what’s the correct endpoint and request format (headers, body, multipart, etc.)?
  3. How can I implement this via an HTTP Request node in n8n — without using the OpenAI/GPT nodes?
2

Hey @kiril_kogan hope all is well.

I don’t believe openrouter support Whisper-style audio transcription natively.

is there a reason why you want to use openrouter and not one of the models directly, for example Gemini?

3

Hi @kiril_kogan, as what I see in the openrouter, OpenRouter doesn’t support whisper from OpenAI or any TTS model. It only used for text-based LLM.

So, the answer is No.

4

Hi @kiril_kogan.
I’ve been working on something similar (albeit not with OpenRouter), and I actually think it is possible to use OpenRouter with Whisper or similar audio transcription. Here’s a github link to a similar project.

The HTTP request setup would probably be pretty simple: I think this is a good starting point but I can’t test it.

Method: POST
URL: openrouter.ai/api/whatevertheactuallinkis
File: Binary (might need to use a conversion node)
Auth: Your API key
5

Hi @kiril_kogan I think OpenRouter supports Whisper: this guy has done it on GitHub:

As for the setup:

Method: Post
URL: Put in your API link
File: Binary
Model: Whatever multimodal model you choose
Auth: Your API key
6

I’ve tried searching the OpenRouter docs, but there isn’t an audio transcription method available via POST. When I try, it gives me a 405 “method undefined” error. So, as far as I can tell, it no longer works.