How to transcribe a WhatsApp voice note on iPhone (for free)

Where the transcript appears (and why it lags)

The transcript isn't automatic for every message: you trigger it voice note by voice note, with a long-press then "Transcribe". The text appears directly under the audio message, in the conversation bubble, and you can hide it again with a tap.

The processing happens entirely on your iPhone, never on WhatsApp's servers. That's why the very first transcript in a language takes a moment: the device has to download the recognition model first. It's also what makes the feature compatible with end-to-end encryption, since the voice note's content never leaves the phone to be transcribed. On this specific point, native transcription is more privacy-friendly than any third-party service, including ours.

The limits of native transcription

It's genuinely useful, but it has three limits worth knowing before you rely on it.

First, availability. You need iOS 16 at minimum for French, and a fairly recent iPhone. The supported languages widen with each version: a dozen from iOS 16, more than twenty from iOS 17. Worth noting, because it surprises a lot of people: on Android the feature also exists but its stable version doesn't handle French yet (it is being tested in the Android beta 2.26.37.5, September 2026).

Next, quality on difficult voice notes. On a long message, recorded in the street, with several people talking or fast delivery, recognition leaves blanks. And they often land on what matters most: proper nouns, numbers, addresses.

Finally, and this is the main limit, you get word for word. A transcribed four-minute voice note is still four minutes of reading, digressions included. Transcription changes the message's format, it doesn't cut the time it costs you. For a short message, it's perfect. For a five-minute monologue where a single sentence concerns you, it doesn't really solve the problem.

Transcribe a long voice note, or get a summary of it

This is exactly where native transcription hits its limit: when the voice note is long, or when you just want to know what's being asked without reading eight paragraphs.

Another approach is not to transcribe word for word, but to summarise. That's what Abrège does: you forward the voice note to the Abrège contact in WhatsApp, like you'd forward it to a friend, and you get back the key points and the action to remember, in a few lines. If you still want the full text, you can switch on full transcription and receive both.

Three practical differences from Apple's feature: you get the gist rather than the whole thing, the names and numbers come through without gaps, and it handles voice notes up to about 25 minutes, on iPhone and Android alike.

In return, let's be transparent: unlike Apple's transcription which stays on your phone, the audio here is transcribed via OpenAI's API, then deleted from our servers with no copy kept. It's a deliberate trade-off, and it's better to know it to choose knowingly. The first 5 summaries are free, no card required.