Built for Apple Intelligence · iOS 26

Just talk.
We'll track it.

Track anything — drinks, sleep, runs, weight. On-device AI parses what you said into beautiful charts over time. No setup. No cloud.

100% on-device
Zero network
No subscription
9:41
Recording…
|
Logged

Say anything quantifiable.

There's no "tracker setup." The first time you mention something, it becomes a tracker. The second time, it builds a chart.

"I had a headache this morning, energy was a 4, slept 7 hours, drank a boba."
headache · event energy · 4/10 sleep · 7 hrs boba · event
"Did 60 pushups, ran a 7:30 mile, weighed 175 pounds."
pushups · 60 mile pace · 7:30 weight · 175 lbs
"Worked 6 hours of deep work, meditated for 10 minutes, felt anxious about a 6."
deep work · 6 hrs meditation · 10 min anxiety · 6/10
How it works

From your mouth to a chart in under 5 seconds.

01

Open the app and talk.

Recording starts the moment you open the app. No menus. No buttons to find. Just talk.

02

Apple Intelligence parses it.

On-device Foundation Models extract what you actually said — multiple trackers, with values, units, and types.

03

Charts build themselves.

Every entry lands in a tracker — old or brand new. The graphs screen tells you the shape of your life over time.

Why it's different

Designed to disappear.

Most tracking apps fail because logging takes longer than the thing you wanted to log. This one is built around a single primitive: your voice.

Voice-first. No taps to set up.

The very first time you say "weight," it becomes a numeric tracker with units. The first time you say "energy was a 4," it becomes a 1–10 scale. The first time you say "headache," it becomes an event counter. Your schema is your speech.

Recording screen with the words 'Did 20 push-ups today' transcribed live
Privacy by architecture

If it never leaves your phone, it can't leak.

The hard thing about a voice-driven tracker is that voice is intimate. Speak To Track is built so that intimacy never has to be trusted to anyone — not to us, not to a cloud, not to a third-party API.

  • Audio is never persisted. Only your final transcript is stored locally.
  • Speech-to-text runs on-device using Apple's Speech framework.
  • Parsing runs on-device via Apple's Foundation Models. No OpenAI, no Anthropic, no anyone.
  • No accounts, no analytics, no telemetry. If you binary-audit the app you'll find no API keys and no remote endpoints.
  • Works in airplane mode — every part of the pipeline.
0
outbound requests
0
accounts required
$0
subscriptions, ever

What you need

iPhone 15 Pro or newer

The app needs Apple Intelligence, which runs on iPhone 15 Pro / Pro Max and the entire iPhone 16 and 17 lines.

iOS 26 or later

Foundation Models — the API that powers on-device parsing — became publicly available in iOS 26.

Mic & speech permission

Granted once on first launch. Both are used only on-device. You can revoke them any time in Settings.

Stop tapping forms.
Start talking.

$0.99 on the App Store. One-time. No subscription, ever.

Download on the App Store