Fact · Dictation on phones and tablets
Apple says that the permission process applies only to speech recognition with SFSpeechRecognizer, and that SpeechAnalyzer transcriber modules do not send the user's voice audio to Apple's servers.
Checked against its source on October 17, 2026
In the source's own words
“This process only applies to speech recognition using ... transcriber modules don’t send audio data of the user’s voice to Apple’s servers.”
Asking Permission to Use Speech Recognition (Speech framework documentation), Apple Inc., n.d., Discussion.
Open the original- What it measures
- Apple's developer guide on asking permission to use speech recognition, read on 2026-10-17. The same page says that speech recognition with SFSpeechRecognizer involves capturing audio of the user's voice and sending that data to Apple's servers for processing.
- What it does not show
- This is Apple's statement about its own frameworks. It is not an independent test of network traffic, and it does not cover third-party speech engines.