Murble
How to

Voice memos to text: convert a recording you already have

The interview from last month, the lecture a friend sent you, the two hour board meeting sitting in Voice Memos. This is how to turn an existing audio file into a transcript you can search, correct and export, without uploading it anywhere.

By Lukasz Majda Updated 25 July 2026

Import and transcribe, step by step

  1. Get the audio file onto the iPhone

    Save it into Files or iCloud Drive, or AirDrop it from a Mac. If the recording is in Apple's Voice Memos app, use the share sheet: Murble's Share Extension accepts supported audio and saves it as a note for processing.

  2. Import it into Murble

    In the library, choose Import audio file and pick the recording. Murble accepts m4a, mp3, wav and the other audio types iOS can read. One file at a time, so a folder of interviews is a repeated action rather than a batch job.

  3. Let it transcribe on the device

    Transcription runs on the phone using Apple's speech recognition or a Whisper model you have downloaded. Nothing is uploaded. A long file keeps processing inside the app while you read other notes, and Murble tells you when the transcript is ready.

    If you want this to work with no connection at all, install an offline model first: see offline transcription on iPhone.

  4. Correct the transcript

    Every line is time aligned. Tap one to jump the audio to that moment, then fix the word that came out wrong. In a conversation, anonymous speaker labels separate the voices, and you can rename, merge or split them.

  5. Export, or clip a passage

    Export the transcript or a summary as PDF, Markdown, RTF, plain text, JSON, SRT or VTT. You can also select a passage and share just that stretch of audio as an m4a clip, which is the fastest way to send somebody the twenty seconds that matter.

Murble transcript screen with time aligned lines, anonymous speaker labels, an audio player and tap to seek
The transcript view of an imported recording. Lines are time aligned, so tapping one moves the audio there, and the currently playing line stays highlighted.

Correcting a transcript is the real work

No speech model gets names, jargon and crosstalk right every time, and a transcript you cannot fix is a transcript you cannot trust. The parts of Murble that matter most for imported files are the boring ones.

After the transcript, optionally

An imported recording gets the same treatment as one you made in the app: a summary in one of four styles, automatically detected action items, and, on Pro, a structured report, translations and chat with the note. All of it runs on the phone. See on device AI note taking.

Export formats and what each is for

The seven export formats, and the job each one is good at.
Format Use it for
PDFSending a finished, readable document to someone who will not edit it
MarkdownDropping into Obsidian, a wiki or a repository
RTFPasting into Pages or Word with the formatting intact
Plain textThe lowest common denominator, and the easiest to script against
JSONKeeping structure and timings for your own tooling
SRTSubtitles for video editors and most players
VTTSubtitles for the web

SRT and VTT carry the line timings, which is why correcting text inline rather than retyping it elsewhere is worth the small effort.

Honest limitations

Sources

Turn the archive into text

Import, transcribe and export are free with no caps, and the audio never leaves the phone to get there.

Download on the App Store

More detail on the Murble landing page.

Keep reading