Does a voice recorder app need microphone access on iPhone?

Every recording app needs the microphone or it cannot exist, so this permission gets granted almost without thinking. The more interesting question is what else a transcribing app is asking for underneath that, and why it is not the same thing.

The permission you would expect anyway

Microphone access gates exactly one thing: whether the app is allowed to receive audio from the phone's microphone at all. Tap record for the first time and iOS interrupts with a system prompt before anything can be captured. Decline it, and the app has no route to audio — not a degraded recording, not a silent file, nothing written to disk at all. There is no way around this and no reason there should be one: an app that captured sound without asking would not be a permission gap, it would be a bug report.

The permission most people do not expect

A voice recorder that also transcribes calls on a second, separate system — Apple's Speech framework — which has its own permission independent of the microphone one. iOS asks for it the first time an app attempts a transcription, with its own prompt and its own line in Settings. Granting microphone access says nothing about this one; an app can have full permission to record and still have no permission to turn what it recorded into text.

The surprising part is that this permission is asked for even when the transcription stays entirely on the device. It would be reasonable to assume a purely local speech-to-text pass needs no more consent than any other on-device processing an app does to your data. Apple treats it differently: the Speech framework requires this authorization for a request regardless of whether that particular request runs the on-device model or gets retried against a server. The permission gates the operation, not the network path underneath it.

Why two permissions for one action

The split reflects that capturing audio and converting it to text are different operations, even when the same app does both. Not every recording app needs speech recognition — plenty just record and play back — so iOS asks for it separately, at the point an app actually tries to use it, rather than folding it into the microphone prompt everyone grants without much thought. That also means you can check, later, which apps asked for which: an app with microphone access and no speech recognition line has, provably, never attempted to transcribe anything.

What happens if you grant one but not the other

Microphone access without speech recognition access is a fully working recorder that cannot transcribe. Recording, playback and file management all work; asking it to turn a clip into text returns an error or nothing, and no audio goes anywhere in the attempt, because the request never gets past the permission check. Speech recognition access without microphone access is not really a state that comes up, since there is nothing to transcribe if the app was never allowed to capture anything. In practice, denying speech recognition is the one worth knowing about — it is easy to tap through absent-mindedly and then not connect a broken transcribe button back to a permission you said no to weeks earlier.

Checking or changing either one later

Both live in Settings → Privacy & Security, as separate sections: Microphone, and Speech Recognition. Each lists every app that has ever asked, with an individual toggle. An app missing from the Speech Recognition list has never requested it — not because it was silently denied, but because nothing in that app calls the API that would trigger the prompt. Turning either toggle off takes effect immediately with no reinstall needed, and turning it back on does too.

How Voice Studio handles it

Voice Studio asks for microphone access to record and, separately, for speech recognition access the first time you transcribe something — the two prompts appear at different points because they gate different actions. Transcription attempts the on-device path first regardless of anything beyond having granted that permission at all; if the local pass comes back empty or errors, it retries with on-device recognition off, which is the point at which audio could reach Apple's speech service. Recording itself never touches that path — it is a microphone, an AAC encoder, and a file, at whichever of the three quality tiers you picked. There is no account, no CloudKit sync and no analytics SDK in the app, so these two permissions are the only ones it has any reason to ask for.

Common questions

Do I have to grant microphone access to use a voice recorder app?

Yes. There is no way to capture audio on iPhone without it — decline the prompt and the app has nothing to record with, not a degraded version of recording.

Can I record without allowing speech recognition?

Yes. Recording, playback and file management do not need it. What breaks is transcription specifically — attempting it without that permission returns an error or nothing, with no audio going anywhere in the attempt.

Does an app need speech recognition permission if its transcription is entirely on-device?

Yes. iOS asks for this permission for any use of the Speech framework, whether a given request runs locally or gets retried against Apple's server. The permission gates the operation, not which path a particular request takes.

Where do I check which apps have microphone or speech recognition access?

Settings → Privacy & Security has a separate section for each — Microphone and Speech Recognition — listing every app that has asked, with a toggle per app.