How can I transcribe Telegram voice messages using Google Gemini in n8n?

Overview
I am building an n8n workflow to transcribe Telegram voice messages using Google Gemini, but I’m running into issues. My workflow needs to:

1. Retrieve the voice file from Telegram
2. Convert the file to a compatible format for Gemini
3. Send the audio file to Gemini for transcription
4. Handle the transcription result

I have attempted to set up an HTTP request but remain unclear on the best approach for these steps. I’ve reviewed documentation and various online resources, yet no clear solution has emerged.

Environment Details

• n8n version: 1.76.0 (Self Hosted)
• Database: Default (SQLite)
• Running n8n via: Docker
• Operating system: Ubuntu

If anyone has implemented this workflow or can offer troubleshooting guidance, code snippets, or best practices, your help would be greatly appreciated.

Thanks in advance!