Skip to content

It's not just an algorithm. It's an algo-rhythm.

The teleprompter that follows your voice

You speak and the text moves with you. Promter detects when you're talking and advances the script at your pace, word by word. And the detection happens entirely inside your device: none of your audio is sent anywhere.

Promter's presentation screen with voice-paced scrolling active: the live mic waveform and the script advancing word by word.

The difference nobody explains properly

Almost every "voice" teleprompter uses speech recognition. It tries to transcribe what you say and match it against the text. It works beautifully in the demo and falls apart in real life, the moment an accent shows up, or street noise, or a brand name, or that technical term the cloud has never heard.

Promter does something else. It doesn't try to understand what you're saying. It senses that you're speaking and calculates, from your measured pace, when each word should end.

The result: there's nothing for an accent to break. There's no word it "doesn't know". And it doesn't need internet, because there's no cloud in the path.

So is there AI in here?

We'd rather be exact about this.

There is a model running inside Promter. It's small, it runs entirely in your browser, and it does exactly one thing: look at the sound coming through the microphone and answer whether that's a human voice. Yes or no, many times a second. The technical name for it is VAD, voice activity detection.

That model doesn't transcribe, doesn't know any language, doesn't know what a word is and never goes anywhere near your script. It can't read.

What decides when the text moves is arithmetic: your measured speed says how long each word should last, and the pointer advances while you're speaking. Deterministic math, no guessing.

And because the VAD runs on your device, none of this needs internet or sends audio to anyone.

Four modes, one icon

ModeWhat happens
Fixed paceScrolls on its own, at the speed you set
Sound-gatedOnly moves while you're speaking
Voice-trackedAdvances and highlights word by word, at your measure
CalibrationMeasures your real speaking speed

You switch between them with one tap, no menu.

There's a 50 to 300 WPM slider always visible, a 3-second countdown before you start, and a live mic waveform so you can see it's hearing you.

What it doesn't do

It doesn't transcribe, doesn't take voice commands and doesn't know when you've improvised. If you go off script and talk for two minutes off the top of your head, the text keeps moving at your pace, because it follows rhythm, not words.

And it needs microphone permission. Without a mic, you use fixed pace mode.

Check out the other features:

Flies by QR Code.

Point one device's camera at the other's screen and the script crosses over. No cable, no cloud, no account, no attachment. Send one script, just your settings, or the entire backup.

Learn more

Offline and Private.

Turn off the wi-fi. It keeps working. Your scripts stay on your device, there's no account, and nothing you write ever leaves.

Learn more

Rich Editor

Just like your content, our editor is RICH when it's time to write. Mark the tone, the emphasis and the pause. It all shows up exactly the same when you present, and one click strips it if you'd rather have plain text. Saves itself, always.

Learn more

Record in up to 4K

Open your camera inside the app. Record video up to 4K at 60 frames per second without looking away from the screen, without stepping out of frame.

Learn more

You're the star.

And the camera operator, the director and the entire crew.

Which is why: a keyboard shortcut for every function, a Bluetooth clicker that works with zero setup, and a screen that won't go dark mid-read.

Learn more

Mirror rig

Horizontal and vertical mirroring, independent, so the reflection lands readable on the glass. Set it once and it remembers.

Learn more

Yes, it's 100% free. No, you're not the product.

Promter runs entirely on your device. No server holding your scripts, no database with your account, no speech recognition API billing by the minute, no AI model billing by the request.

Works with pretty much everything.

Mac, PC, Android, iOS, Linux. All you need is an up-to-date browser. No app store, no installer, no waiting on a download.

Made it this far down? You probably have questions.

So I put together this FAQ to set your mind at ease.