Speak with a glance.
Swar turns a glance into full sentences, then says them out loud. Built for people who have lost speech and fine motor control.
- Camera video stays
on your device - English, Spanish,
French, Vietnamese
Swar
GazeWhat is this about?
Thinking of what you might mean…
Pick what you meant to say:
Look at a tile. It fills while you hold your gaze, then it is chosen.
The problem
Letter by letter is not conversation
Most assistive devices make you pick every single letter through a slow channel like eye movement. That is barely fast enough to say "I'm in pain". Swar skips the spelling: you give a small hint, and it offers whole sentences to choose from.
5 to 15
words per minute with letter-by-letter tools
2 to 3
whole sentences offered at a time, to pick from
4
languages: English, Spanish, French and Vietnamese
0
seconds of camera video ever uploaded
Letter by letter
0.0s
Still spelling…
Swar
0.0s
An illustration, not a benchmark. Letter-by-letter tools typically manage 5 to 15 words per minute, and this shows one short phrase at about 10.
How it works
Four steps from a glance to a voice
Each step uses the same simple input, so there is nothing new to learn along the way.
Step 1
Give a hint
Look at a topic tile, such as food, pain or family. One look is all it takes.
Step 2
Get full sentences
The app suggests a few complete sentences. It uses what it knows about you, like names and routines, so they fit you.
Step 3
Confirm or try again
Pick the sentence you meant, or ask for new ones. Nothing is spoken until you choose.
Step 4
Hear it spoken
The sentence plays out loud right away, so the conversation keeps moving.
Languages
Speaks in the language they think in
Suggestions are written, and spoken aloud, in the language you choose.
English
I'd like some water, please.
Español
Quisiera un poco de agua, por favor.
Français
Je voudrais de l'eau, s'il vous plaît.
Tiếng Việt
Cho tôi xin một ít nước.
Who it's for
Made for people who can't use a keyboard
Built for people with
- ALS
- Brainstem stroke
- High-level spinal cord injury
- Late-stage muscular dystrophy
These conditions can leave eye movement or a blink as the one reliable way to communicate.
Not the right tool if you can type
This is not a general text-to-speech app. If you can use a keyboard, a touchscreen, or a mouse, any ordinary text-to-speech app will serve you better.
Typing exists here only as a backup for the moment the camera can't track you. It was never meant to be the normal way to use this.
For families and caregivers
You set it up once, they use it every day
Most people using this can't operate the setup themselves. That is where you come in.
Set up the camera
Sit the person facing the screen with light on their face. A short calibration teaches the app where they are looking.
Add what makes them them
Enter names, routines, and phrases they use. Suggested sentences then sound like the person, not a stranger.
Questions
Things families ask
Do I need special hardware?
No. It works with the webcam on a laptop or tablet. A dedicated eye tracker would be more precise, and the app is built so one could be added later.
Does my camera video get sent anywhere?
No. Your video is processed inside your browser and is never uploaded. The first time you use gaze mode, your browser downloads the face-tracking model files, so that first run needs internet.
What if it suggests the wrong sentence?
Choose "None of these" to get new sentences, or go back and pick a different topic. Nothing is spoken until you choose a sentence.
Do I need an account?
Yes, a free account. It keeps your names, routines and phrases saved, so a helper can set things up once. Each account only sees its own details.
Let's get started
See it for yourself
Create an account, add a few details about the person, then pick a topic, choose a sentence and hear it.