title: Customizing the coach voice description: How the system-voice coach works, picking a persona on Pro, and tuning speech.

Customizing the coach voice

The coach voice is what makes a workout feel like a workout instead of a glorified timer. All coaching speech is generated on-device by iOS system voices — fully offline, no network needed — with coach personas as a Pro option on top.

How the voice is picked

Everyone. AVSpeechSynthesizer running on-device. The engine picks the best available voice on your phone: "Premium" or "Enhanced" voices if you've downloaded any, otherwise the default. You can download better system voices in iOS Settings → Accessibility → Read & Speak → Voices → English (called Spoken Content on older iOS versions) — RhythmFit notices and uses them automatically. This is the single best upgrade you can make to how the coach sounds, on Free or Pro.

Pro — Coach Personas. Three coaches, each mapped to a matching system voice on your device:

| Persona | Vibe | Best for | | -------- | ------------------------------------- | ----------------------------------- | | Max | Energetic, confident male | HIIT, Tabata, strength sets | | Luna | Warm, calm female | Yoga, breathwork, meditation | | Kai | Steady, friendly male | Runs, cycling, cadence work |

Auto (the default) picks the right coach for each workout type — Max for HIIT, Luna for yoga, Kai for runs. Free users hear the default system voice until they upgrade.

The countdown voice matches the persona — so a Max workout counts down with the same voice it cues with.

Picking a persona

Settings → Coach. Free users see all rows but only System is selectable; the three Pro personas show a lock. Selecting a coach speaks a short sample immediately, so you hear exactly what that persona sounds like on your device. The info button on any row opens a description with a Hear Sample button.

Because personas are backed by system voices, they sound best when the matching premium voices are installed. If a persona's preferred voice isn't on your device, RhythmFit walks down a candidate list and finally falls back to the best default voice — you always hear something sensible, offline, with zero lag. When that fallback makes a persona sound identical to System, the detail sheet shows a hint naming the voice to download (e.g. "Tom" for Max) in iOS Settings → Accessibility → Read & Speak → Voices → English (older iOS: Spoken Content).

Tuning the speech timing

Settings → Voice Coach has a speech rate slider (0.5× to 1.5×). Slower is calmer; faster cues cleaner but can blur on long lines. Default 1.0×.

Language and accent

iOS system voices are localized — pick a voice from the right language in iOS Settings and RhythmFit follows. The cue text stays English (RhythmFit ships English-only for now), so if your device language is set to something else the pronunciation can sound odd. The cleanest fix is to keep your iOS device language as English while we work on localization.

Disabling the coach entirely

For users who want only the metronome (cadence drills, racewalking, silent runs), turn off the Voice Coach toggle in Settings. The metronome keeps running; you lose pre-block notes, in-block reminders, the countdown read, and the celebration audio. The engine still progresses normally — it just doesn't talk.

This survives across sessions until you turn it back on.