Interview transcription that never leaves your iPhone

LoroNote transcribes interviews entirely on your iPhone or iPad, so a source’s voice is never uploaded to a transcription server. Speakers are separated and labelled on the device, so a two-person interview comes back split into interviewer and subject, and a one-hour recording finishes in roughly three minutes on an iPhone 15 Pro.

What makes interview audio different

Two voices that have to stay apart

A transcript that runs both people together is close to useless for pulling quotes. LoroNote identifies speakers on the device, and renaming one applies it across the whole transcript — the labels stay attached through TXT and SRT export.

The recording is somebody else’s risk

A source is trusting you with their voice, not a vendor. Audio that is never uploaded cannot be exposed in that vendor’s breach, handed over under a third-party subpoena, or retained after you delete your copy.

They are long, and they happen where signal is not

Forty-five to ninety minutes is ordinary, and interviews get recorded in cafés, cars, and corridors. Cloud plans meter exactly this: Otter Pro allows 1,200 minutes a month and 90 minutes per conversation. LoroNote caps neither, and works with the network off.

Names are the words that matter most

A generic model will mis-spell the surname, company, or place your piece is about — the exact words you cannot fix by skim-reading. The custom dictionary lets you fix a spelling once so it comes out right in later recordings.

How an interview goes through it

  1. 1 Record in the app, or import the file from a dedicated recorder — MP3, WAV, M4A, MP4, and MOV all come in as they are.
  2. 2 Transcription runs on the device when the recording ends. Airplane mode changes nothing about the result.
  3. 3 Speakers arrive already separated. Rename “Speaker 1” to your subject once and it applies throughout.
  4. 4 Checking a quote: tap the line and the audio jumps to that second, so you verify against the tape rather than your memory of it.
  5. 5 Export TXT for writing, or SRT if the interview is going into video with timed captions.

Why this matters for source protection

Cloud transcription makes your source’s voice a record held by a company they never agreed to deal with — reachable by that company’s staff, its breaches, and its legal obligations. On-device transcription removes that party from the picture entirely: LoroNote operates no transcription server, so there is no copy anywhere to reach. It is a property of how the app is built, not a promise in a policy that can be revised.

What it will not do for you

The parts that are still your job:

Common questions

Can LoroNote separate the interviewer from the subject?

Yes. Speaker identification runs on the device and labels each part of the transcript automatically. You can rename a speaker once to apply it everywhere, reassign any segment by hand, and the labels are preserved when you export or share.

Can I import interviews from a dedicated recorder?

Yes. Import MP3, WAV, M4A, MP4, or MOV files from a field recorder, a phone, or anywhere else, and LoroNote transcribes them on your device exactly as it does its own recordings.

Is it safe to transcribe interviews with confidential sources?

The audio is never uploaded, because LoroNote runs no transcription server — so there is no third-party copy to breach or subpoena. That removes the largest technical risk. It does not settle your own obligations to the source, or the recording-consent rules where you are.

How long can an interview be?

No plan caps the length or the number of recordings. Practical limits are your device’s storage and battery. Processing runs at roughly 19x real time on an iPhone 15 Pro, so a 90-minute interview takes about four and a half minutes.

Related

Try it on one interview

LoroNote is a free download. Transcribe a real interview with the network off and check the speaker labels against the tape before you move any work across.

Download LoroNote