Why do so many workout apps stop the user's music when they speak a cue?

Because the platform default does exactly that, and so does the obvious fix done halfway. Apple documents that the default iOS audio session silences other background audio when the app plays, and that the playback category, which keeps cues alive when the screen locks, is nonmixable by default. On Android, a cue that requests full AUDIOFOCUS_GAIN instead of transient ducking focus tells the music app it has lost focus, and on Android 12 and higher Google documents a forced fade-out that leaves it muted. Both platforms have a documented way to lower the music for a few seconds and give it back; the pages here are about using it.

Is a spoken cue or a haptic better for telling someone a workout interval has ended?

Use both, and design for the moment one of them fails. Spoken cues carry information a tap cannot, such as what comes next and how long it lasts, but the audio channel is shared with the user's music and can be taken away by a call, by Siri or by an earbud falling out. A haptic reaches the person even then, but only if the device is touching them. The accessibility section covers haptics as the second channel in detail, and these pages route a cue to a haptic whenever the documented signals say speech should not play.

What does this section not cover, and why?

Cross-platform frameworks and anything we could not read. The React Native, Expo, Flutter and Capacitor documentation sites were unreachable from our research environment, so these pages make no claims about audio plugins for those frameworks. We also publish no measured latency, ducking depth or battery figures, because we have not run those tests, and where Apple's or Google's pages are silent, for example on which volume stream each Android usage maps to, the page says so rather than guessing.

Audio Cues & Ducking for Workout Apps

Audio Coaching

A workout app’s voice has to share the audio channel with the user’s own music, and most audio bugs in fitness apps come from getting that sharing wrong: the cue stops the music and never gives it back, goes silent when the screen locks, talks over a podcast, carries on through the speaker after an earbud falls out, or fails without a sound in the background on a newer Android release. Each of those is a documented platform behaviour with an exact API name attached — duckOthers and notifyOthersOnDeactivation on iOS, AUDIOFOCUS_GAIN_TRANSIENT_MAY_DUCK and AUDIOFOCUS_REQUEST_FAILED on Android — and these pages are organised around those names. Every claim comes from Apple’s or Google’s own developer documentation, read on 3 October 2026.

This is engineering guidance for the audio layer only. Using haptics as the second channel when the audio is busy is covered in accessibility, and running the workout session itself on a watch is covered in watch apps. The product-level requirement, duck the user’s music instead of stopping it, is set out in the HIIT app guide; these pages are how to meet it.

Three evidence rules hold throughout. The documentation sites for React Native, Expo, Flutter and Capacitor were unreachable from our research environment, so nothing here describes a cross-platform audio plugin. We publish no latency, ducking-depth or battery numbers of our own; the one ducking figure quoted is Google’s, and attributed to it. And where the platform pages are silent — whether an iPhone app keeps running between cues when nothing is playing, which volume stream an Android usage maps to, whether TextToSpeech manages audio focus — the page says it could not verify the point rather than filling the gap. Store policy, including App Store guideline 2.5.4 on background services, belongs to compliance.

Configure the session (iOS)

The category that survives the lock, the option that lowers the music instead of stopping it, and the speech synthesizer that has to be held on to.

When something else happens

A phone call, an earbud falling out, the screen locking. The transitions where workout audio actually breaks.

Android audio focus and speech

Asking for focus the right way, the Android 15 rule that fails background cues silently, TextToSpeech, and the usage constant nobody reads.

On the wrist, and proving it

Speaker or Bluetooth on Wear OS and Apple Watch, and a device test matrix built from documented events.

Frequently asked questions

Why do so many workout apps stop the user's music when they speak a cue?: Because the platform default does exactly that, and so does the obvious fix done halfway. Apple documents that the default iOS audio session silences other background audio when the app plays, and that the playback category, which keeps cues alive when the screen locks, is nonmixable by default. On Android, a cue that requests full AUDIOFOCUS_GAIN instead of transient ducking focus tells the music app it has lost focus, and on Android 12 and higher Google documents a forced fade-out that leaves it muted. Both platforms have a documented way to lower the music for a few seconds and give it back; the pages here are about using it.
Is a spoken cue or a haptic better for telling someone a workout interval has ended?: Use both, and design for the moment one of them fails. Spoken cues carry information a tap cannot, such as what comes next and how long it lasts, but the audio channel is shared with the user's music and can be taken away by a call, by Siri or by an earbud falling out. A haptic reaches the person even then, but only if the device is touching them. The accessibility section covers haptics as the second channel in detail, and these pages route a cue to a haptic whenever the documented signals say speech should not play.
What does this section not cover, and why?: Cross-platform frameworks and anything we could not read. The React Native, Expo, Flutter and Capacitor documentation sites were unreachable from our research environment, so these pages make no claims about audio plugins for those frameworks. We also publish no measured latency, ducking depth or battery figures, because we have not run those tests, and where Apple's or Google's pages are silent, for example on which volume stream each Android usage maps to, the page says so rather than guessing.

All 42 questions in Audio Coaching, answered

Independent comparison, last reviewed October 3, 2026. Pricing, rate limits, and feature availability change often — confirm current details in each provider’s official documentation before you commit. Product and company names are trademarks of their respective owners; AIFitnessAPI is not affiliated with, endorsed by, or sponsored by any product listed here.