Can’t process videos larger than 25MB for transcription – tried MP4→MP3 conversion (CloudConvert fails due to memory)

Questions
1

Hey everyone :waving_hand:

I’m trying to build an automated transcription workflow in n8n using OpenAI’s Whisper API.

My goal:

  1. Watch a Google Drive folder for new video uploads

  2. Convert the uploaded .mp4 video to .mp3

  3. Send the audio file to OpenAI Whisper for transcription

The problem is that any file larger than 25MB breaks the flow — I can’t upload or pass it to Whisper due to API limits.
I’ve tried multiple workarounds but can’t get past the size/memory issues.

What is the error message (if any)?

  • CloudConvert → “Memory error” (fails when converting large videos)

  • OpenAI Whisper → rejects files >25MB (limit error)

What I’ve tried so far

:white_check_mark: Converting MP4 → MP3 using CloudConvert and FreeConvert APIs — fails for big files.
:white_check_mark: Attempted to compress audio before sending — but FFmpeg isn’t available natively on n8n Cloud.
:white_check_mark: Tried uploading directly to Whisper — rejected because of size.

What I’m trying to achieve

  • Extract audio from videos using FFmpeg or any working n8n approach.

  • Compress it enough to stay under 25MB without losing quality.

  • If the file is still large, automatically split it into chunks and send each piece to Whisper for transcription.

Basically, I just want a reliable way to handle larger video/audio files inside n8n before sending them to Whisper.

Workflow snippet

(Simplified overview)
Google Drive Trigger → Download File → Convert to MP3 → Send to Whisper

Information on your n8n setup

  • n8n version: (please fill in)

  • Database: Default (SQLite)

  • EXECUTIONS_PROCESS: Default (own/main)

  • Running via: n8n Cloud

  • Operating System: —

Question

:backhand_index_pointing_right: How can I process or compress large videos inside n8n (Cloud) before sending to OpenAI Whisper, given that FFmpeg isn’t built in and external APIs fail on big files?
Would love guidance on how others handle this — e.g. using self-hosted FFmpeg, chunking audio, or any alternative node/workaround.

2

The OpenAI Whisper API itself has a 25 MB file upload limit when used directly through the API or tools like ChatGPT’s “transcribe audio” feature. This limit is set by OpenAI’s servers for performance and bandwidth reasons it’s not a technical limitation of the Whisper model itself, but rather a safeguard for the API service.

3

any other solution, i just need to compress the video, so that i can do it

4

Try the convert.online site.