Voice activated teleprompter that follows you
Read at your own pace. The text moves as you speak, waits when you stop, and finds its place again if you skip or ad-lib. Free in your browser, with no app and no account.
110 words · about 47 sec at 140 wpm
Opens full screen. Space pauses, arrow keys change speed, Esc closes. Nothing is uploaded — your script stays in this browser.
Keyboard shortcuts
| Space · K · B | Play / pause |
|---|---|
| ↑ / ↓ | Faster / slower (10 wpm) |
| ← / → | Nudge back / forward one line |
| Page Up / Page Down | Nudge three lines (clickers, foot pedals) |
| + / − | Bigger / smaller text |
| M | Mirror |
| F | Full screen |
| R · Home | Back to start |
| Esc | Close |
Wheel or drag to move the text by hand. Tap the text to pause.
How to use the voice teleprompter
-
Paste your script
Paste it into the editor or open a .txt file. Put anything you do not plan to say, like camera cues, in [square brackets].
-
Pick your language
Auto picks the language from your script. If your accent differs, choose it yourself, for example English (India) or Português (Portugal). Matching the accent improves recognition.
-
Start and allow the microphone
Your browser asks for mic access the first time. A Listening label appears while the prompter is following you.
-
Just read
The text moves as you speak. Stop to think and it stops. Skip a sentence and it jumps ahead once it is sure where you are.
-
Fix things by hand if needed
If it loses you, press ← or → to move a line, or drag the text. It picks up from the new spot.
How voice scroll works
The teleprompter uses your browser speech recognition to turn what you say into text, word by word, as you speak. It then compares the last few words it heard with the script around your current position and moves the reading line to match.
Speech recognition is never perfect, so the matching is forgiving. A misheard word, an extra "um" or a small change in wording does not throw it off. If you skip a few words or a sentence, it finds the matching words further along and catches up.
Big jumps need confirmation. If what it heard could match a spot far from where you are, it waits for a few more matching words before moving there. That stops a common phrase from sending the text to the wrong paragraph.
Words in [square brackets] are ignored by voice scroll, so you can keep stage directions in the script without the prompter waiting for you to say them. Voice scroll is in beta. For scripted reads at a steady pace, the regular timed mode on the free teleprompter is still the most predictable.
Voice scroll or timed scroll?
Both modes are in the same tool. Switch between them in the settings or the control bar.
Use voice scroll when
You pause a lot, you ad-lib between lines, your pace changes through the script, or you are working alone and cannot reach the keyboard. It also helps when you are new to reading a prompter and a fixed speed feels like it is chasing you.
Use timed scroll when
The room is noisy, you are on Firefox, you read at a steady pace, or the script is full of names, jargon or numbers that speech recognition may mishear. A fixed wpm is also easier to rehearse to.
Combine with recording
Voice scroll works on its own page and in the main teleprompter. If you also record yourself, test a short take first to check that the microphone is working for both.
Use with a rig or on a call
Voice scroll does not care where the screen is. It works with mirrored text in a glass rig and in a narrow window under your webcam during a call.
Which browsers support voice scroll?
Voice scroll depends on the speech recognition built into your browser.
| Browser | Voice scroll | Notes |
|---|---|---|
| Chrome (desktop, Android) | Yes | Works best. |
| Edge (desktop) | Yes | Works best, along with Chrome. |
| Safari (Mac, iPhone, iPad) | Yes | Works. Allow the microphone when asked. |
| Firefox | No | No speech recognition. Use timed scrolling instead. |
If voice scroll is not available, the teleprompter falls back to timed scrolling at your chosen speed.
Privacy and languages
Where your audio goes
Speech recognition is done by your browser vendor speech service: Google for Chrome, Apple for Safari, Microsoft for Edge. Your audio is not sent to Short.now.
Your script stays local
The script is saved only in this browser. Matching your words against it happens on your device.
30+ languages, picked for you
The language is detected from your script, so a Russian or Portuguese script is heard in Russian or Portuguese. You can also pick it yourself: English, Spanish, Portuguese, French, German, Italian, Dutch, Polish, Russian, Ukrainian, Turkish, Swedish, Greek, Arabic, Hebrew, Hindi, Indonesian, Vietnamese, Thai, Japanese, Korean and Chinese, with regional accents.
Free, no account
Many teleprompter apps put voice scrolling behind a subscription. Here it is part of the free tool, with no account.
Tips to make voice scroll more accurate
- Record in a quiet room. Music, a TV or other people talking confuse recognition more than anything else.
- Use a decent mic. A headset or USB mic close to your mouth works better than a laptop mic across the desk.
- Speak clearly at your normal pace. You do not need to slow down, but mumbled words are harder to match.
- Keep cues in brackets. [pause], [smile] or [cut to demo] stay visible but the prompter will not wait for them.
- Pick the right accent. English (India) or English (UK) can recognize you better than English (US) if that is how you speak.
- Recording yourself too? The video teleprompter records your camera while you read.
Recorded a long video? Get the Shorts out of it.
Short.now finds the best moments in your video, reframes them to 9:16 and adds captions, so one recording becomes a week of Shorts, Reels and TikToks.
Questions
A teleprompter that listens to you and scrolls the script to match what you are saying. It stops when you pause and moves on when you continue, so you do not have to set a speed.
Yes, this one. Voice scroll is included in the free browser teleprompter with no sign-up or subscription.
Yes, in Safari. Chrome and Edge on a computer usually give the smoothest results.
Firefox does not include the speech recognition feature that voice scroll relies on. Use Chrome, Edge or Safari, or use timed scrolling in Firefox.
No. Your browser sends audio to its own speech service (Google, Apple or Microsoft) to turn it into text. Short.now does not receive your audio.
The text waits while you ad-lib. When you return to the script, it finds the words you are saying and continues from there.
Yes. It works in over 30 languages and accents, including Spanish, Portuguese, French, German, Russian, Ukrainian, Arabic, Hindi, Japanese, Korean and Chinese. It detects the language from your script, or you can choose it.
