Mermer
iPhone
A voice recorder that writes down what was said.
Recordings are hard to use. You end up with a list of grey rectangles labelled by date, and the only way to find one sentence is to listen to all of them again.
Mermer transcribes your recordings on your iPhone, works out who was speaking, and makes the words searchable. No account, no server, no subscription, and Mermer never uploads your recordings for transcription or speaker recognition.
Rich 0:04
So we said we'd decide about the garden fence before the weekend.
Anna 0:11
Right, and the quote was eight hundred, which felt steep for six panels.
Rich 0:19
Let's get one more quote. If it's close, we go with the first one.
Record, then read it back
There is nothing to set up and nothing to sign into. Tap the button and talk.
-
Tap record
Recording keeps going when the screen locks, so you can put the phone down mid-conversation without losing the rest of it.
-
The transcript appears
When you stop, Mermer transcribes the audio using Apple's on-device speech model, when your language is supported. Processing time varies with the length of the recording, and it lays the words out as paragraphs you can actually read.
-
Search it, share it
Find any recording by a phrase from inside it. Share the audio file, or the transcript as text, anywhere iOS can share.
It knows who was talking
A two-person conversation is unreadable as one wall of text. Mermer separates the voices and labels each turn.
Speakers, separated
Mermer works out how many people are talking and splits the transcript into turns — down to the word, so a quick back-and-forth doesn't get stitched onto the wrong person.
Name them once
Tap "Speaker 1" and type a name. It applies across the whole transcript, and to anything you share.
Remembered voices
Choose to remember a voice and Mermer recognises it in later recordings, labelling those turns by name automatically. Recognition is approximate, so it's worth checking a name before you rely on it, and it's best to ask the person before you remember their voice. Voices you remember are listed on one screen and can be deleted whenever you like.
Search what was said
Type a phrase and Mermer shows the recordings containing it, with the matching line quoted so you know why it matched.
Timestamps when you want them
Turn timestamps on to see when each turn happened, or leave them off and read it like a page of dialogue.
Undo a deletion
Delete a recording and an Undo button stays on screen. Change your mind and the recording and its transcript come straight back.
Processed on your phone, not someone else's
This is the part most transcription apps cannot offer, because their transcription happens on someone else's computer.
No cloud processing
Mermer includes no advertising or analytics SDKs. Speaker separation uses a bundled, open-source, on-device library (FluidAudio), with its models shipped inside the app. There's no cloud transcription and no developer-operated server: Mermer does not upload your recordings or transcripts for processing.
No account, no server
Nothing to sign up for, nothing to log into, and no database somewhere with your conversations in it.
No analytics
No tracking, no telemetry, and no crash-reporting or analytics service is integrated. Mermer does keep a small troubleshooting log on your device to help fix bugs; it stays local and is never uploaded. The App Store privacy label reads "Data Not Collected".
Voices stay yours
The profile stores numerical voice characteristics, not a playable recording. Delete it and it is removed from your device, though names already saved in past transcripts stay, and the recordings themselves are separate.
Free, with nothing to buy
No subscription, no in-app purchases, no ads, no per-minute transcription limit. There is no payment code in the app at all.
Deleting a recording
Deleting a recording removes its audio and transcript from Mermer. Undo stays on screen briefly, and the item is cleared from Mermer's trash the next time you open the app. Copies you've shared, and copies in existing backups, are separate.
About the speech model. A model download may be needed. iOS manages speech model installation and updates, not Mermer, and that download comes from Apple to your phone; no audio is uploaded, then or ever. Mermer shows the progress plainly instead of pretending nothing is happening.
Coming to the App Store
Mermer is on its way. In the meantime, the support page explains how it works, and the privacy policy sets out exactly what happens to your recordings, which Mermer processes on your device and never uploads for transcription or speaker recognition.