Can you export a voice memo before transcribing it?

It is a reasonable thing to assume: transcribing feels like the step that finishes a recording, so exporting before it happens can feel like jumping ahead. The two are not actually in a sequence. They are two independent actions that both happen to start from the same recording, and only one of them has any reason to care whether the other has run yet.

What exists right after you stop recording

The moment a recording ends, there is exactly one thing sitting in the library: an audio file. Transcribing is a step you take afterward, deliberately — it does not run automatically the instant you press stop, and a recording with no transcript yet is not a broken or unfinished recording, it is just a recording nobody has asked to turn into text. Nothing about the file itself is waiting on that to happen.

Exporting the audio never checks for a transcript

Sharing a recording as its .m4a is a file operation — copy this file to wherever you are sending it. There is no step in that process that looks at whether a transcript exists, because a transcript is not part of what makes an audio file an audio file. You can record something and share it thirty seconds later, before you have even opened the recording again, and nothing about the export is incomplete for having skipped transcription. The words simply are not part of what is being sent.

This is genuinely useful in the ordinary case where the audio is the urgent part: sending a voice message to someone right now, handing off a clip to an editor who needs it today, or getting a recording off the phone before you have decided whether you personally need a transcript of it at all. None of that has to wait on a speech recognizer finishing anything.

Exporting the transcript is the one export that actually needs one

A TXT or JSON transcript export is a file built from text, and text has to exist before it can go into a file. If a recording has not been transcribed, there is nothing yet to write into either format — not an error, just an absence, the same way a document with no words typed into it has nothing to save. Transcribing first and exporting the transcript second is not a rule the app enforces so much as the only order that makes sense for what that export is actually doing.

That is the real distinction hiding inside the original question. "Export" sounds like one action with one set of requirements, but a voice recorder with both an audio export and a transcript export is really offering two exports of two different things, and only one of them has a transcript as an ingredient.

Where a full backup fits in

A full backup is one JSON file covering the whole library — every recording's title, date, duration and transcript, gathered into a single record you would use to restore the library rather than to read casually. Running that backup does not wait around for you to finish transcribing everything first. A recording you have not gotten to yet is included with whatever it already has — its title, its date, its duration — and no transcript text, because none has been produced. That is not a gap in the backup; it is an accurate record of where that recording actually stands right now. Transcribe it later and the next backup reflects that.

Why it is worth keeping the two independent

Forcing transcription before export would tie a fast, offline, entirely local action — copying a file — to a slower one that sometimes involves Apple's speech service and always takes at least a little processing time. There is no good reason for the first to wait on the second. A recorder that let you get audio out immediately and figure out the words on your own schedule is doing the more useful thing, especially for anything time-sensitive where the audio itself is what somebody needs, not a transcript of it.

What transcribing afterward does not change

If you already exported the audio and transcribe the recording later, nothing about the file you already sent changes or needs to be redone — it was never missing anything, because it was never supposed to carry the transcript in the first place. Getting the words to that same person later is a separate, ordinary export: TXT, JSON or a copied block of text, sent whenever the transcript is actually ready, not a correction to something you already handed off.

How Voice Studio handles this

Recording, transcription and export are three separate, unrestricted features in the free app, and none of them gates another. Share a recording's .m4a the moment it is captured, whether or not you have ever opened it again to transcribe it. Exporting the transcript as TXT or JSON, or copying it, works once a transcript exists — produced by tapping to transcribe, which always attempts the on-device path first and only retries against Apple's server if that local pass comes back empty or errors. The full backup in Settings picks up whatever each recording currently has, transcript included only where one has actually been made, so an untranscribed backlog never blocks a backup from running today.

Common questions

Do you have to transcribe a recording before you can share the audio?

No. Sharing the .m4a is a plain file operation that does not check whether a transcript exists. You can export the audio the moment a recording finishes, whether or not it has ever been transcribed.

What happens if you try to export a transcript for a recording that has not been transcribed yet?

There is nothing yet to put in the file, since transcribing is a step you take yourself rather than something that happens automatically when you stop recording. Transcribe the recording first, then export the transcript.

Does exporting the audio trigger transcription in the background?

No. Exporting just copies the .m4a to wherever you send it. It does not call the speech recognizer, and transcribing later does not alter the audio file you already exported.

If you transcribe a recording after already exporting the audio, do you need to redo that export?

No. The audio file was complete and unaffected either way. Getting the words out afterward is a separate export of the transcript, done whenever it is ready, not a re-export of the audio.