Frequently asked questions
Why does the app need an iPhone 15 Pro or newer?
The parsing brain is Apple's on-device Foundation Models framework. That framework only ships on Apple-Intelligence-capable devices: iPhone 15 Pro / Pro Max and the entire iPhone 16 and 17 lines, plus M-series iPads. The App Store hides Speak To Track from devices that can't run it, so you'll only see it if your iPhone is compatible.
Does anything I say leave my phone?
No. Audio is never persisted — not even temporarily on the device. Speech-to-text runs on-device via Apple's Speech framework. Parsing runs on-device via Apple's Foundation Models. There are no API keys and no remote endpoints in the binary. The only network feature is when you manually email yourself a CSV export from Settings.
The record screen says "Apple Intelligence is still preparing." What should I do?
On first enable, iOS downloads the Apple Intelligence model — it's several gigabytes and only runs while the phone is on Wi-Fi, charging, and idle. Leave the phone plugged in on Wi-Fi for 10–30 minutes and try again. You can check the status in Settings → Apple Intelligence & Siri.
How do I edit a tracker or fix a wrong entry?
There's no tap-to-edit UI by design. From the trackers screen, tap Talk to AI at the bottom and just say what you want changed — e.g. "the headache I just logged should be a migraine", "delete my last energy entry", or "merge sinus pressure into headache."
Can I export my data?
Yes. Settings → Export data generates three CSVs — entries, trackers, and raw voice logs — and opens an email pre-filled to your address with the files attached. Your data is yours.
How much does it cost?
$0.99 on the App Store — one time. No subscription, no in-app purchases, no ads, no analytics, no upsells. You pay once and you're done. Apple's on-device model means there's no ongoing inference cost to pass on to you, which is what makes a one-time price possible at all.
Does it work offline?
Completely. Every part of the pipeline — recording, transcription, parsing, charting — runs locally. The app works in airplane mode.
How do I delete all my data?
Settings → Clear all data. There's a confirmation dialog. This wipes every tracker, entry, and voice log from the device. There's nothing on a server to delete because there's no server.
My speech got mis-transcribed. What can I do?
Apple's on-device Speech framework occasionally mis-hears things, especially proper nouns or unusual unit phrases. Speak To Track shows the live transcript while recording, so you can stop and re-record. If the entry already saved with the wrong word, use Talk to AI to fix it by voice.
Where can I read the privacy policy?
The short version is on the
main page. The literal version: the app has no networking entitlements beyond what
MFMailComposeViewController requires for sending mail. Audio is never persisted. Transcripts and parsed entries live only in SwiftData on your device. There are no analytics, no crash reporters with PII, and no third-party SDKs that phone home. If you want to verify this, the bundle is small enough to inspect.
Will there be iPad, Mac, or Apple Watch versions?
iPad runs the iPhone build in compatibility mode today. A native iPad layout, Mac Catalyst, and an Apple Watch capture complication are all on the v2 wish list — but v1 is intentionally iPhone-only to keep the surface small.
Can I sync across devices?
Not in v1. The app is single-device and local-first on purpose. iCloud sync via SwiftData's CloudKit integration is being considered for a future version — it would still be end-to-end Apple-managed, with no backend on our side.