newaiforyou

Log in to continue

Please continue by logging in.

Subvocal

Smart glasses from The Subvocal Company that read silently mouthed words with radio sensing and type them on your computer, now on preorder.

Visit SubvocalUpdated Nov 3, 2026
Subvocal homepage screenshot

Subvocal Introduction

Subvocal is a pair of smart glasses that lets you type by mouthing words without making a sound. Antennas in the frame send radio waves, similar to Bluetooth, through the lower face, and models trained by the company decode how the tongue, lips and soft palate move into text, which the site says reaches more than 200 words per minute. It does not use cameras and does not read involuntary thoughts: according to the makers, you start input with a trigger key on the computer or phone you are typing into. Decoding runs in the cloud through a companion app on your laptop or phone, usually in under 500 ms. Subvocal is built as an input method for the apps you already use, such as ChatGPT, Claude, Cursor, Slack and email. It is made by Aarush and Arnav of The Subvocal Company, a Y Combinator company in San Francisco, who launched it on Hacker News in November 2026. The glasses are sold by preorder in the United States only, with shipping estimated for the first quarter of 2027.

Subvocal key facts

LaunchedNov 2, 2026 on Hacker News
PricingPaid, from $499
Free tierNo, hardware sold by preorder
PlatformsSmart glasses with a companion app for laptop or phone
Sign-upPreorder with a $199 refundable deposit (US only)
AI modelsOwn models trained on radio sensing data
Made byAarush and Arnav, The Subvocal Company (YC), San Francisco

How we checked Subvocal

Not tested hands-on · Checked on Nov 3, 2026

Subvocal is hardware that has not shipped yet, so we could not try it. We read the website, the preorder page and the preorder terms, and the makers’ Hacker News launch post with their replies to commenters.

Subvocal Key Features

Silent Text Input

Mouth words without making a sound and the glasses turn them into text, at more than 200 words per minute according to the site.

Radio Sensing, No Cameras

Antennas in the frame use radio waves that pass through the face to follow the tongue, lips and soft palate. There is no camera and no lip reading.

Works in Your Existing Apps

Subvocal is built as an input method, so text goes wherever you would normally type, such as ChatGPT, Claude, Cursor, Slack, email and documents.

Short Personal Calibration

A brief calibration when you first set up the glasses adapts the model to the way you articulate words.

Background Noise Does Not Matter

Because it does not listen to your voice, traffic, music and nearby conversations do not get in the way of input.

Subvocal Use Cases

Office workers: dictate messages, notes and prompts for coding agents without disturbing the people at the next desk.

Commuters: write on a train or in a cafe without being overheard.

People in loud places: enter text where a microphone would pick up too much noise, such as a workshop or a busy street.

Subvocal User Guides

  1. Reserve a pair on the Subvocal website with a refundable $199 deposit (US delivery addresses only).

  2. Pay the remaining $300 of the $499 Early Bird price when the company asks for the final payment.

  3. When the glasses arrive, pair them with the companion app on your laptop or phone and complete the short calibration.

  4. Press the trigger key on your device and silently mouth the words you want to type.

Subvocal Pricing

Preorder only: a $199 refundable deposit now, credited toward the $499 Early Bird price (regular price shown as $699), with $300 due later. Taxes, shipping and optional prescription lenses cost extra.

Subvocal launch and community

Nov 2, 2026 on Hacker News · 698 points, 340 comments

“We’re building smart glasses that let you talk to your computer without talking out loud. You can mouth words silently and discreetly and the glasses decode them to text at the speed of speech.”
From the maker's launch post

What people asked

Does it work in loud places?

Yes. The maker said you can dictate silently where it is noisy, and that a later software update may use the radio signal to clean up the audio when you do speak out loud.

Can it attach to my existing prescription glasses?

No. The makers considered a clip-on module but chose their own frames so the antennas sit in the same place for everyone. Prescription lenses are an extra-cost option.

Is there a wake word?

No. The maker said input starts with a trigger key on the device you are typing into, not with a wake word.

Read the launch post on Hacker News

Subvocal limitations

  • The glasses have not shipped; the company estimates the first quarter of 2027 and says the date is not guaranteed.
  • Preorders are open only to delivery addresses in the United States.
  • Decoding needs the companion app and runs in the cloud, according to the makers.
  • Each user needs a short calibration, and the company says accuracy is still improving.

Subvocal Frequently Asked Questions

Is Subvocal free?
No. Subvocal is hardware sold by preorder: a $199 refundable deposit now, credited toward the $499 Early Bird price (listed at $699). Taxes, shipping and prescription lenses cost extra.
When does Subvocal ship?
The company estimates the first shipments in the first quarter of 2027 and says this is not a guaranteed date. Delivery is also subject to FCC approval.
Does Subvocal use a camera?
No. It uses radio sensing built into the glasses, not cameras or lip reading, and it does not need your webcam.
Where does the decoding happen?
The makers say decoding runs in the cloud through a companion app on your laptop or phone, with end-to-end latency usually under 500 ms.
Can I preorder outside the United States?
Not yet. Preorders are open only to customers with a delivery address in the United States.

Subvocal alternatives