Audio to text on iPhone and iPad, without uploading anything
Every route on this page ends the same way: the file stays on your iPhone or iPad, Whisper Large V3 Turbo transcribes it locally in over 100 languages, and you export TXT or an SRT with timecodes and speaker labels. There is no upload, no account, and no monthly minute allowance to spend. Start from the format you actually have.
Start from your file format
| Format | Usually comes from | Guide |
|---|---|---|
| M4A | Voice Memos, Zoom, QuickTime, GarageBand | M4A to text |
| MP3 | Podcasts, voice recorders, webinar downloads | MP3 to text |
| WAV | Field recorders, Audacity, dictation hardware | WAV to text |
| MP4 / MOV | Camera footage, screen and meeting recordings | MP4 / MOV to text |
Or start from the job
- Voice memo to text Where Apple’s built-in transcript stops — languages, speakers, subtitles — and what to do then.
- Subtitle generator (SRT) Timed .srt files from audio or video, with speaker labels, generated on the device.
- Timestamped transcripts Sentence-level timecodes, tap-to-replay, and SRT export that keeps the timing.
The workflow, whatever the format
- 1
Import the file
Tap Import File and choose an MP3, WAV, M4A, MP4, or MOV from Files or iCloud Drive. Video works directly — LoroNote reads the audio track — so nothing has to be converted first.
- 2
Confirm the language
The spoken language is detected automatically from over 100. Names, product codes, and jargon go in the custom dictionary beforehand so they come back spelled correctly.
- 3
Transcribe on the device
Whisper Large V3 Turbo runs on the device GPU, so airplane mode changes nothing. Keep the app open while a long file processes.
- 4
Review the speakers
Speakers are separated and labeled on the device. Rename "Speaker 1" once and every one of their lines updates; reassign any segment that was labeled wrong.
- 5
Export TXT or SRT
TXT for the words, SRT when the timecodes need to travel with them. Speaker labels survive both, and on devices with Apple Intelligence you can also generate a summary and action items first.
How long it takes
Processing time tracks the length of the recording, not the size of the file — which is why a 600 MB WAV and a 40 MB MP3 of the same hour finish at the same time.
| Recording length | Time to transcribe |
|---|---|
| 5 minutes | about 16 seconds |
| 30 minutes | about 1.5 minutes |
| 1 hour | about 3 minutes |
| 3 hours | about 10 minutes |
Measured by LoroNote on an iPhone 15 Pro at roughly 19x real time. Newer devices transcribe faster; older supported models take longer.
What this does not do
Stating the limits up front saves you the download if one of them is a dealbreaker:
- It transcribes files, not links — there is nowhere to paste a YouTube or Vimeo URL.
- It transcribes; it does not translate. Whisper Large V3 Turbo returns the language that was spoken.
- Export is TXT and SRT. No VTT, no ASS, and no captions burned into the video itself.
- It runs on iPhone, iPad, and Apple Watch only — there is no Android, Windows, Mac, or web version.
- On-device transcription needs iOS 18.6 or later and iPhone 13 or later, iPhone SE (3rd generation), or a supported recent iPad.
Common questions
Which audio and video formats can LoroNote transcribe?
MP3, WAV, M4A, MP4, and MOV are the formats the App Store listing names, which covers essentially everything an iPhone, a voice recorder, a meeting app, or a video editor produces. For video, only the audio track is used.
Does converting audio to text on iPhone need an internet connection?
Not with LoroNote. Whisper Large V3 Turbo runs on the device, so import and transcription both work in airplane mode. That is also the simplest way to verify any app is genuinely local rather than quietly uploading.
Is there a limit on file length or how many files I can transcribe?
No plan cap on minutes or files. Cloud services meter minutes because every minute costs them GPU time on a server; transcribing on your own device has no per-minute cost to recover. The real constraints are storage, battery, and processing time.
How long does transcription take?
About 3 minutes for a one-hour recording on an iPhone 15 Pro, measured at roughly 19x real time. Newer devices are faster and older supported models slower, and the length of the recording matters far more than the size of the file.
What can I export?
TXT and SRT. SRT carries standard hh:mm:ss,mmm timecodes and keeps speaker labels, so it works both as a subtitle track and as a timed transcript. There is no VTT or ASS export, and captions are not burned into video.
Which devices can run on-device transcription?
iOS or iPadOS 18.6 or later on iPhone 13 or later, iPhone SE (3rd generation), iPad mini (6th generation) or later, iPad Air with M1 or later, iPad Pro with M1 or later, or iPad (11th generation) or later. There is also an Apple Watch app for recording.
Try it on one file
LoroNote is a free download. Import a single recording, put the device in airplane mode, and watch the transcript appear with the network off.
Download LoroNote