Most AI transcription apps make you open a separate app, wait for an upload, then hunt down the text afterward. This shortcut skips all of that: share an audio clip or record one on the spot, and OpenAI’s Whisper model hands back a clean transcript in seconds, right inside whatever app you were already using.
What it actually does
Point OpenAI: Transcribe at an audio file, or use it to record a new clip, and it sends the audio to OpenAI’s Whisper transcription model and returns text. It pays particular attention to WhatsApp: if the audio it receives is a WhatsApp voice note, the shortcut recognizes the source, transcribes it, and reopens WhatsApp so you can paste the result straight into a reply. Anything else you share, from a Voice Memo to a podcast clip, just comes back as plain text you can copy, save, or hand to another shortcut.
Apps you need
- Shortcuts — built into iOS 16 and later, no separate install required.
- An OpenAI account with an API key — the shortcut calls OpenAI’s transcription endpoint directly, so this is non-negotiable.
- WhatsApp (optional) — only relevant when the audio you share originated as a WhatsApp voice note; every other source works without it.
Getting it set up
- Tap Add Shortcut on this page to open it in the Shortcuts app.
- Scroll the preview to the bottom and tap Add Shortcut to confirm.
- Run it once and paste in your OpenAI API key when it asks. As of version 13, that prompt fires before anything else, so a missing or invalid key stops you before you’ve wasted time recording.
First-run setup
Grab the API key from your OpenAI developer account before you start; the shortcut has nowhere else to pull one from. Head to your account’s API keys page, generate one, and paste it in when the shortcut asks. Every transcription after that draws on your OpenAI account’s usage balance rather than any subscription tied to the shortcut itself, so cost tracks how much audio you actually send it.
Running the shortcut
Day to day, there are two ways in. Share an audio file to the shortcut from the Share Sheet, whether that’s a Voice Memo, a downloaded clip, or a WhatsApp voice note, and it transcribes it automatically, reopening WhatsApp first if that’s where the audio came from. Launch OpenAI: Transcribe on its own instead, and it records fresh audio right there, no separate voice memo step needed. Version 13 switched the shortcut to OpenAI’s newer gpt-transcribe model and turned off the old forced-English setting, so it should follow other languages more naturally than earlier versions did.
FAQ
Does OpenAI: Transcribe cost anything to use?
The shortcut itself is free to download, but the transcription isn’t. It calls the OpenAI API using your own key, so usage draws from your OpenAI account’s balance rather than anything charged by the shortcut.
Will it always reopen WhatsApp after it transcribes something?
Only when the audio actually came from a WhatsApp voice note; the shortcut checks the source before deciding what to do next. Share a Voice Memo or any other audio file instead, and it just leaves you with the text, no app-switching involved.
Can other shortcuts use the transcript this one produces?
They can. A 2025 update changed the output format so OpenAI: Transcribe passes text back in a form other shortcuts can grab directly, instead of just displaying it on screen.
Tips
Since its first release in December 2024, the author has quietly swapped auth methods and transcription models under the hood several times, most recently moving to OpenAI’s gpt-transcribe model, without changing how the shortcut looks to run. None of that changes how you use it day to day, but it does mean accuracy has improved release over release.
