iMessage Voice Message Transcription: What Works, What Doesn't, and a Better Way

Apple added automatic voice message transcription to iMessage in iOS 17. It was a welcome change -- finally, you could read audio messages instead of listening to them. But anyone who uses it regularly knows it's far from perfect.

Here's what you need to know about iMessage audio transcription, where it falls short, and what to do when you need something more reliable.

How iMessage transcription works

Starting with iOS 17, iMessage automatically transcribes audio messages and displays the text directly below the voice note. There's no setting to toggle -- it just happens. Both the sender and recipient need to be on iOS 17 or later for the transcription to appear.

The feature works on-device, meaning your audio isn't sent to Apple's servers. That's great for privacy. But it also means the quality of the transcription depends entirely on your device's processing power and Apple's local speech model.

The problems with iMessage transcription

Users have reported several consistent issues:

Accuracy drops with longer messages. Short, clearly spoken voice notes transcribe reasonably well. But anything over 30 seconds -- especially with natural speech patterns, pauses, or filler words -- tends to produce garbled results.

Background noise kills it. If the person recording the message is in a car, on the street, or in a busy room, the transcription can be unusable.

No summary or highlights. Even when the transcription is accurate, you get a raw wall of text. For a long message, you still have to read the whole thing to find the one piece of information you actually need.

No way to turn it off (or fix it). Many users have complained that there's no setting to disable transcription, and no way to correct errors. You're stuck with whatever the system generates.

Limited language support. The feature works primarily in English and a handful of other languages. If the voice message is in an unsupported language, you get nothing.

What about saving to Voice Memos?

One workaround is to save the audio message to Voice Memos (long-press the message, tap "Save to Voice Memos"), then use Voice Memos' transcription feature (available in newer iOS versions). This gives you a slightly different transcription engine, but it's an extra step and the accuracy improvement is marginal.

You can also export the audio file and upload it to a third-party transcription service, but at that point you're spending more time managing files than you would have spent just listening to the message.

A simpler approach

Yadda works with iMessage voice notes just like it works with any other voice message. Forward the audio message to (570) YADDA-ME, and you'll get a text back with both a full transcription and a short summary highlighting the key points.

The advantage over Apple's built-in transcription: better accuracy (it uses more powerful AI models than what runs locally on your phone), a summary so you don't have to read the whole thing, and it works regardless of what iOS version the sender is on.

When to use what

Apple's built-in transcription is fine for short, clearly recorded messages in English where you just need a rough idea of what was said.

Yadda is better when the message is long, the audio quality is iffy, you need a summary, or you want a reliable transcription you can reference later.

Either way, the days of having to find a quiet spot and headphones just to hear what your friend said are over. Try Yadda the next time someone sends you a voice note you can't listen to right now.