How to Transcribe Audio to Text on iPhone (4 Ways)

The short answer: If you are recording something new, Apple’s Voice Memos will transcribe it for free on iOS 18 and later. If you have an audio file already, or you need to know who said what, you need an app that accepts imports and labels speakers.

“Transcribe audio to text on iPhone” covers four genuinely different jobs, and picking the wrong method is why people conclude it does not work. Here is each one, what it is good at, and where it stops.

1. Apple’s built-in transcript — free, for recordings you make

On iOS 18 and later, open a recording in Voice Memos, and you can view a transcript of it and search the text. Nothing to install, nothing to pay, no upload.

This is the right first stop, and for a lot of people it is the last one. It handles a single voice well: your own notes, a lecture, a phone-side dictation of a to-do list.

Where it stops:

• It only knows about recordings inside Voice Memos — an audio file someone sent you is out of scope.
• There are no speaker labels, so a conversation arrives as one continuous block of text.
• There is no summary, and no way to hand the text off in a structured form.

We walked through the full flow, including exporting the text, in our guide to transcribing voice memos on iPhone.

2. Dictation — when you want text, not a recording

If the goal is words on a screen rather than an archived recording, skip transcription entirely. Tap the microphone on the keyboard in Notes, Messages, or any text field and talk. The text appears as you speak, and you can say “new paragraph”, “comma”, and “question mark” to punctuate.

This is the fastest route to a written note, and it is free. It is also the wrong tool for anything you did not plan: you have to be speaking to the phone, in real time, and there is no audio kept afterwards to check against.

3. An app that transcribes recordings and imported files

This is what you need when the audio already exists, or when more than one person is talking. Record and Transcribe records like any recorder, then transcribes automatically once you stop — and it accepts files you share into it, so old recordings, exported voice notes, and meeting audio all work the same way.

Two things it adds that the built-in options do not:

• Speaker labels, worked out from the audio itself, with no voice training or setup. In a meeting or interview this is the difference between a transcript you can use and one you have to replay.
• Reliability on long sessions. Audio is written in checksummed segments as you record, which matters on the hour-long recordings where a single file is most likely to be lost. We wrote about why iPhone recordings get lost after it happened to us.

Recording is free and unlimited, three transcriptions are included so you can judge the accuracy on your own audio, and summaries and longer memos are the paid part. The honest boundary: if you only ever record yourself for thirty seconds at a time, Apple’s free transcript is enough and we will not pretend otherwise — that comparison is laid out in full on our Voice Memos comparison.

4. Free web tools — with one caveat

There are browser-based transcribers you can upload a file to from Safari. They work, and for a podcast clip or something already public they are perfectly reasonable.

The caveat is worth stating plainly: uploading audio to a website means handing a recording of someone’s voice to a company whose retention policy you have not read. For a client conversation, a medical appointment, an interview under embargo, or anything involving a third party who has not agreed to it, that is a bad trade. Keeping audio on the device exists precisely for this reason.

Which one to use

Your situation	Use
Recording yourself, want it free	Voice Memos transcript
Want text immediately, no recording needed	Dictation from the keyboard
Audio file you already have	An app that accepts imports
A voice message or WhatsApp voice note	The built-in transcripts, or save the file
Two or more people talking	An app with speaker labels
Hour-long session you cannot afford to lose	An app that writes audio in segments
Non-sensitive clip, one-off	A free web tool

Getting better accuracy, whichever you pick

Transcription quality is set almost entirely by the recording, not the software. Four things do most of the work:

• Put the phone on the table, screen up, roughly equidistant from the people talking. Not in a bag, not in a pocket, nothing resting on it.
• Avoid speakerphone audio. Room speakers reflecting off hard walls degrade badly; a phone recording of a phone call is the worst case.
• Let people finish. Heavy crosstalk is the single biggest cause of both wrong words and wrong speaker labels.
• Say unusual names once, clearly, early. Names, drug names, and product names are what transcripts get wrong most, and a clear first mention gives the model something to anchor to.

The realistic expectation

Good audio of clear speech gets you a transcript that needs light editing, not a rewrite. Names, jargon, and anything said over the top of someone else will need fixing. Nothing on this page — free or paid, Apple’s or ours — removes the final read-through, and any tool promising otherwise is selling you something.

If you want to test the difference on your own audio rather than take our word for it, the free tier includes three transcriptions and needs no card.