What happens if you deny speech recognition permission on iPhone

Open a recording app for the first time and iOS usually asks for two things, one after the other: permission to use the microphone, and permission to use speech recognition. They look like a formality and get tapped through without much thought, but they gate two genuinely different capabilities, and knowing which is which explains most of what people run into later.

The two prompts, and what each one actually gates

Microphone

This is the permission that lets the app capture sound at all. Deny it, and there is no recording — the app cannot get audio from the hardware, full stop. Every recorder app needs this one; there is no version of "record audio" that works around it.

Speech Recognition

This is a separate permission that lets the app hand audio to Apple’s Speech framework and get text back. It has nothing to do with whether the microphone works. Deny it, and recording is completely unaffected — you can still record for as long as you like, pause, resume, save the file. What you lose is the ability to turn that audio into a transcript, because the system will not run the recognizer for an app without this permission.

Why iOS treats these as two separate decisions

Recording audio and transcribing it are different acts with different implications, and Apple’s permission model reflects that rather than bundling them. You might be comfortable with an app hearing you live while you talk into it, and less comfortable with that same audio being run through a recognition model — or the other way around. Splitting the prompt means you get to make that call once per app, not have it decided for you by whichever permission happened to be asked first.

What granting speech recognition does not mean

It is easy to read "speech recognition" as "sending my voice to a server", and sometimes that is exactly what it means — but not always, and not automatically. Apple’s Speech framework can run entirely on the phone for languages that have an on-device model installed, and apps that use it typically try that path first. Granting the permission only lets the recognizer run at all; it says nothing about which of the two paths — local or Apple’s server — actually handled a given piece of audio. That is a separate question, answered by language coverage and whether the on-device model is present, not by which permission you tapped.

If you said no and want to change it

Nothing about the decision is permanent, and nothing you recorded while it was off is lost. Go to Settings → Privacy & Security → Speech Recognition, find the app in the list, and turn it on. The permission is checked at the moment you actually ask for a transcript, not at the moment you recorded — so audio you captured before granting the permission can still be transcribed afterwards, the same as anything else. The Microphone entry lives one screen up, under Settings → Privacy & Security → Microphone, in case that one needs the same fix.

How Voice Studio handles it

Voice Studio asks for microphone access to record and, separately, for speech recognition access to transcribe — the second is only requested when you actually try to transcribe something, not on first launch. Denying it does not touch recording in any way; you keep every feature that involves capturing and managing audio, and you can grant the permission later from Settings whenever you decide you want transcripts. When it is granted, transcription always attempts the on-device path first and only retries against Apple’s server if that local pass comes back empty or errors — that fallback is about the audio and the language, not about which permission got you into the recognizer in the first place. Voice Studio has no account and no server of its own either way.

Common questions

Does denying speech recognition permission stop my recordings from working?

No. Recording, playback, markers, search and everything else that does not require turning audio into text keeps working exactly as before. Only the transcript is affected.

Does granting speech recognition permission mean my audio is always sent to a server?

No. The permission only allows the recognizer to run at all. Whether a given recording is handled on-device or sent to Apple’s server depends on language coverage and whether the on-device model is installed, not on the permission itself.

If I turn the permission on later, can I still transcribe recordings I made while it was off?

Yes. The permission is checked when you ask for a transcript, not when you recorded the audio, so nothing you captured earlier is excluded once you grant it.

Where do I find the speech recognition permission for a specific app?

Settings → Privacy & Security → Speech Recognition lists every app that has asked for it, with a toggle per app. Microphone access is a separate entry one screen up in the same Privacy & Security menu.