How to Transcribe Voice Memos, Voicemails and Interview Recordings on iPhone, Offline (No Subscription)
Beyond chat apps: how to turn an iPhone Voice Memo, a saved voicemail or a long interview recording into text on the phone itself, with nothing uploaded. Which formats work, what to expect from long files, how to get a cleaner transcript, and what the one-time unlock covers.
Most people meet voice transcription through a chat app: a friend sends a three-minute voice note and you want to read it instead. But the recordings that really need transcribing are usually the ones you made yourself. The lecture you recorded in Voice Memos. The voicemail from the clinic with a date in it. The forty-minute interview for a dissertation or an article. The idea you dictated in the car and now cannot remember. This post is about turning those into text on an iPhone without a subscription and without sending the audio to anyone's server.
The tool is Voice Note to Text, a small iPhone app whose speech-recognition model ships inside it. It was built for chat voice notes, but it opens any .m4a, .mp3, .wav or .opus file, which is what makes it useful for everything below. It is iOS-only, free to download with three transcriptions included, and a single one-time purchase of $2.99 unlocks unlimited use.
Voice Memos: the two-tap route
Apple's Voice Memos app records in .m4a, a format the transcriber opens directly. Open Voice Memos, tap the recording, tap the three-dot menu, choose Share, and pick Copy to Voice to Text from the share sheet. The app decodes the file, loads its model on first use, and the transcript appears under the file name a few seconds later for a short memo. Tap Copy to paste it into Notes or an email, or Share .txt to save it as a file.
If the memo is already in the Files app, or came from a different recorder, open the transcriber directly and use Pick a voice note & transcribe to browse to it. Both routes produce the same result; the share sheet is just faster when you are already looking at the recording.
Voicemail: getting the audio out first
On carriers that support Visual Voicemail, the Phone app lets you share a voicemail as an audio file: open the voicemail, tap the share button, and the same Copy to Voice to Text option appears. The result is a text version of the message with the appointment time, the reference number or the callback number in it, which is far easier to act on than replaying the audio three times with your finger over the pause button.
One caveat worth knowing: voicemail audio is heavily compressed and often recorded on a poor line, so expect a few more errors than with a clean memo. Numbers and names are the usual casualties. For anything important, check the digits against the audio once; the transcript gets you to the right place in seconds, and the recording confirms it.
Interviews and lectures: long files
This is where most phone transcription apps either stop, cap you at a few minutes, or start charging per minute. There is no length limit here. A ten-minute recording is typically transcribed in well under a minute on a recent iPhone, and longer files take proportionally longer, so a one-hour interview is a few minutes of waiting rather than an upload, a queue and an email. Plug the phone in, leave it on the screen, and do something else.
For interviews there are two practical tips. First, record with the phone closer to the quieter speaker; a transcript is only as good as the audio, and the interviewer is nearly always louder than the interviewee. Second, transcribe the raw file before you trim or edit it, so the text and the audio line up when you go back to check a quote. The transcript does not include timestamps or speaker labels; it is a plain text of what was said, which is what most people want to search, quote from and summarise.
For students, the same workflow turns a recorded lecture into searchable notes. Ask permission to record, transcribe the .m4a afterwards, and paste the text into whatever you use for revision. The model handles around 100 languages and detects the language automatically, so a lecture in Spanish or Hindi needs no setting changed.
Why offline matters for your own recordings
A chat voice note is usually trivial. A recorded interview, a medical voicemail or a dictated draft is not: it may contain someone else's personal details, an unpublished piece of work, or something you were told in confidence. Cloud transcription services process audio on their servers under terms that vary from vendor to vendor. Here the model runs on the iPhone and the app makes no network requests at all. You can verify this yourself by switching on airplane mode before transcribing: it still works. There is no account to create and the App Store privacy label reads Data Not Collected.
The trade-off is the download size. Because the speech model is inside the app, the install is around 96 MB. That is the price of the first transcription being instant and offline, and of never having to wait for a model to download later.
Getting a cleaner transcript
- Reduce background noise at the source. A recording made in a café transcribes worse than one made in a car with the windows up. Nothing fixes this afterwards.
- Speak in full sentences when dictating. The model punctuates from the rhythm of your speech; a list of fragments comes out as a run-on.
- Keep one language per recording where you can. Detection is per file, so a memo that switches languages halfway through will be transcribed in whichever one dominates.
- Convert first only if you must. The app opens .m4a, .mp3, .wav, .opus, .ogg, .aac, .caf, .aiff, .flac, .amr and .3gp directly, so there is normally no reason to run a file through a converter.
What the free tier and the unlock cover
The first three transcriptions are free, regardless of length, so you can test it on a real interview rather than a ten-second sample. After that, the one-time $2.99 purchase removes the limit. There are no minutes, credits or monthly plans, and no difference in quality between the free and unlocked transcriptions. If your need is a single long recording, the three free runs may be all you ever use.
The app runs on any iPhone with iOS 15.1 or later. If you would rather listen than read, or you need an audio file in a format another device can play, the companion Opus to MP3 converter covers the other direction; and if your interest is chat voice notes specifically, the original guide to transcribing WhatsApp voice messages on iPhone walks through that case step by step.
The apps mentioned here are independent utilities built by the author. They are not affiliated with, endorsed by, or connected to WhatsApp LLC or Meta Platforms, Inc. WhatsApp is a registered trademark of Meta Platforms, Inc.