Voice Productivity · iOS · watchOS · Web
A note-taking app for people who think out loud.
Ideas never arrive when catching them is convenient. They show up mid-walk, on a drive, or at 3 a.m. when the closest device is a watch. Unlock, find the app, type, decide where it belongs. Every step is trivial on its own. Stacked together, they are enough for the idea to be gone.
The thought is rarely one clean note anyway. It comes out as a tangle of errands, deadlines, and half-formed ideas. I wanted to say all of it at once and have the app work out the structure on its own.
It turns unstructured speech into organized to-dos, reminders, and notes, waiting on your phone, your wrist, your home screen, and the web. You speak. It sorts. It is there when you go looking.
A solo project. I defined the product, designed the interface and the brand, and built the iPhone and watchOS apps, the widget, the web client, and the Supabase backend. It is primarily native Swift rather than my usual React Native, because the watch, the widget, and Apple's platform-level speech are not things you reach comfortably from anywhere else.
You never pick a format first. Say "schedule the accountant call, do the laundry, and call Mom tonight," and the app splits the stream afterward. A task becomes a to-do, anything with a time becomes a reminder, a thought worth keeping becomes a note.
Sorting will never be perfect, so nothing it produces is locked. Rewrite it, change the category, split it, reschedule it, or check it off. The goal was never an infallible parser. It was making the fix cheaper than typing the thought yourself.
An early build showed the transcript only after you stopped talking. It captured everything and still felt unreliable. A waveform can only say something is happening. Words landing as you speak them are what say I understood you.
Typing runs slower than the thought and pushes you into editing an idea before it has finished forming. Speaking keeps it loose. This is not a replacement for long-form writing. It shortens the distance between having an idea and having it somewhere safe.
Everything lands in the same Supabase account and surfaces everywhere else immediately. A to-do caught on the watch is on the phone and the web before you reach a desk. You never have to remember which device caught the thought.
To-dos, reminders, and notes stay separate enough to scan without becoming three different products, and the interface leads with what is still active. It stays narrow on purpose. It is not a second brain, a project manager, and a calendar in a trench coat.
The watch covers the moments nothing else can. Mid-walk, mid-workout, or with the phone in another room, you raise your wrist, say the thing, and it has synced before you get back. Not a compressed iPhone app: it captures fast and confirms the thought is safe.
At a desk, reaching for a phone breaks your focus. The bigger canvas also makes the web the best place to work through a backlog and clear out what is already done. Capture happens everywhere else. This is where it gets room to breathe.
Opening the app is friction too, so the widget puts recording on the home screen. The gap between three taps and one sounds trivial until the thought disappears somewhere around tap two.
Certain phrases map to actions: start a list, push something to tomorrow, mark a task done. The vocabulary is deliberately small, because a known language beats pretending every imaginable instruction will work.
The name does two jobs. Speakeasy is the promise of the interaction, where speaking is the easy part. It also points at a 1920s New York speakeasy, and the mark follows it there: it reads as a sound wave first, but its silhouette is a city skyline against a dark night.
SpeakEasy Notes shipped on the App Store on July 13, 2026.
The part I care about more is that it became a tool I actually use. It came out of a frustration I kept running into rather than a market gap I went looking for, and using it daily keeps turning up things worth sharpening. The product stays small on purpose. Its value is not how many modes it has. It is how quickly it finishes one job.
SpeakEasy taught me that removing a single second from a workflow can matter enormously when that second sits between a person and a thought that is about to vanish.
It also pushed me well outside my usual React Native habits, into native Swift, watchOS, widgets, streaming speech, and the problem of designing one coherent product across four surfaces at once.
The best description of it is still the simplest. I talk. It listens. It sorts. And the thought is still there when I need it.