Imagine travelling on a crowded Mumbai local train and needing to reply to an urgent message. Typing is difficult, while speaking your message aloud is not ideal. This is where Subvocal AI glasses and silent speech to text could help.
Subvocal AI glasses from The Subvocal Company are built for exactly that moment. You mouth words without making a sound, and the glasses convert your silent speech to text at more than 200 words per minute.
The startup has raised $1.3 million from Y Combinator and Afore Capital, and pre-orders opened on October 30, 2026. Below is what the technology does, how it works, what it costs and where it stands for buyers in India.
Table of contents
- What Is Subvocal? The Subvocal Company Explained
- What Is Subvocalization? A Quick Primer on Silent Speech
- How Subvocal Silent Speech Recognition Works
- Subvocal Typing Speed: How Fast Is 200 WPM?
- Subvocal Specifications and Key Details
- Subvocal Price and Pre-Order Details
- Subvocal Price in India and Availability
- Who Is Backing Subvocal? Y Combinator and Afore Capital
- Open Questions Around Subvocal AI
- AI Smart Glasses You Can Buy in India Right Now
- FAQs on Subvocal and Silent Speech Technology
- The Bigger Picture for Silent Speech Interfaces
What Is Subvocal? The Subvocal Company Explained
Subvocal is a pair of glasses designed to do one job well: let you enter text into your computer without a keyboard and without speaking aloud. Messages, notes, long-form writing and AI chatbot prompts are all intended uses.
The company positions it as an input method, sitting alongside your keyboard and voice dictation. Whatever app is open on your screen, the words simply land there as you silently articulate them.
The founders are Aarush Yadav (CEO) and Arnav Kumar, both computer science students at the Georgia Institute of Technology. They built the first prototype at Stanford’s TreeHacks hackathon in February 2026 using muscle-sensing electrodes and machine learning.
Although they did not win the hackathon, they moved to San Francisco and started building the product full time, which is how Subvocal YC ended up in the Fall 2026 batch of Y Combinator.
What Is Subvocalization? A Quick Primer on Silent Speech
Subvocalization is the tiny, mostly unconscious movement of your tongue, jaw and throat muscles when you “say” words in your head, for example while reading a book quietly. Your speech system partly fires even when no sound comes out.
A silent speech interface (SSI) is any device that picks up those faint signals and turns them into text or commands. Researchers have chased this idea for years, usually with electrodes taped to the face and neck. MIT’s AlterEgo headset from 2018 is one well-known early example.
Subvocal works with deliberate silent articulation. You consciously mouth the words, just without voicing them. The company states clearly that the glasses do not read arbitrary or involuntary thoughts.
How Subvocal Silent Speech Recognition Works
According to the company, the glasses send radar signals through your face. When you silently form words, small movements in your speech muscles change how those signals bounce back. An AI model then reads those changes and turns them into text.
Radar sensing here means the frame emits very low-power radio waves and measures the reflections. A car’s parking sensor works on a similar principle to detect an obstacle.
Yadav describes it as tracking internal speech muscles through non-contact sensing, so nothing sticks to your skin.
Three design choices define the product:
- No camera. Subvocal says it does not use or require cameras for silent speech input, so this is not lip reading.
- Opt-in by design. If you are not deliberately trying to say something through Subvocal, there is nothing for it to transcribe.
- Personal calibration. A short setup session teaches the system how your face moves when you speak silently.
The company is refreshingly upfront about mistakes. Its site admits the system will occasionally get words wrong, so the software leans on context and easy corrections, much like autocorrect on your phone.
Subvocal Typing Speed: How Fast Is 200 WPM?
WPM stands for words per minute, the standard way to measure how quickly someone enters text. Subvocal says its prototypes have already crossed 200 WPM. Here is how that compares with everyday input methods.
| Input method | Typical speed | Silent | Private in a crowd |
|---|---|---|---|
| Phone thumb typing | Around 35 to 40 WPM | Yes | Yes |
| Keyboard typing | Around 40 WPM (average user) | Mostly | Yes |
| Voice dictation | Around 130 to 160 WPM | No | No |
| Subvocal (company claim) | 200+ WPM | Yes | Yes |
That last row is the whole pitch. Voice dictation tools are already fast, but nobody wants to read their emails aloud in an open office or a library. Subvocal promises dictation speed with keyboard-level privacy.
One caveat matters a lot. The 200 WPM figure comes from the company itself and has not been independently tested. Subvocal has also not published any accuracy numbers yet, only saying accuracy keeps improving as it collects more data and trains larger models.
Subvocal Specifications and Key Details
| Detail | Subvocal glasses |
|---|---|
| Product type | Smart glasses for silent text input |
| Sensing technology | Radar based, non-contact |
| Camera | None |
| Claimed speed | 200+ words per minute |
| First-time setup | Short personal calibration |
| Intended uses | Messages, notes, writing, AI assistants |
| Prescription lenses | Available at extra cost |
| Early bird price | $499 (US) |
| Reservation deposit | $199, refundable |
| First shipments | Q1 2027 (estimated) |
| Final hardware specs | Still in development |
Battery life, weight, supported operating systems and connectivity details have not been announced. The company says its current images show the design direction, and the final hardware may change.
Subvocal Price and Pre-Order Details
Reserving a pair takes three steps on the official Subvocal pre-order page. You pay a $199 refundable deposit, choose your fit and lens options, and then wait for the first shipment.
The early bird price is $499, so the remaining $300 is due later. Prescription lenses cost extra. On a straight currency conversion, $499 works out to a little over Rs 40,000, before shipping, customs duty and GST.
Subvocal Price in India and Availability
Reservations are currently open in the US only. There is no India launch date, no Indian pricing and no local retail partner announced so far.
If Subvocal does reach India later, import duties and taxes would likely push the final price above the dollar conversion. Early adopters here will also want proof that the system handles Indian English accents and Hinglish well before spending that kind of money.
Who Is Backing Subvocal? Y Combinator and Afore Capital
Y Combinator (YC) is the Silicon Valley startup accelerator behind companies such as Airbnb, Stripe and Dropbox. Getting into a YC batch gives a young company funding, mentorship and a large investor network.
Subvocal’s money came in a pre-seed round, the earliest stage of startup funding, usually raised before a product is finished.
Afore Capital led with $500,000 in April 2026 when the team only had a prototype. YC added another $500,000, and angel investors made up the rest of the roughly $1.3 million total.
The startup also won YC’s Group 1 Product Showcase, which batchmates and alumni vote on. A live demo at the YC office drew a strong reaction from Y Combinator’s David Lieb, who called it the most impressive demo he had seen.
Open Questions Around Subvocal AI
Excitement is fair, but a pre-order product deserves a few honest question marks.
Accuracy in the real world. A demo in a quiet room is one thing. Long dictation sessions, unfamiliar names and technical words are a much tougher test, and no accuracy data is public yet.
Accents and languages. Nothing has been said about support beyond English. For Indian users juggling English, Hindi and regional languages, this will be a big deciding factor.
Software rivals. Within hours of the funding news, a developer shared a free open-source version that uses AirPods sensors, which he said took nine minutes to build.
Another team claims silent speech input on regular laptops and phones without extra hardware. Dedicated glasses will need to prove they are clearly better.
Privacy of body data. Radar readings of your facial muscles are personal data. Buyers will want clear answers on whether processing happens on the device or in the cloud, and how long any data is stored.
AI Smart Glasses You Can Buy in India Right Now
Subvocal will not be on Indian shelves any time soon. If you want smart eyewear today, these options cover very different budgets and use cases.
| Glasses | Price in India | Camera | Standout feature | Buy link |
|---|---|---|---|---|
| Ray-Ban Meta (Gen 2) | From Rs 39,900 (MRP) | 12MP | Meta AI, 3K video, Hindi support | Check today’s deal |
| Pebble Seeq | MRP Rs 12,999 | 8MP | Live translation in 75 languages | Grab it on Flipkart |
| B by Lenskart | Varies by frame | Yes | Google Gemini, UPI payments | Browse Lenskart smart glasses |
| Other AI smart glasses | Budget to premium | Varies | Calls, music, voice assistant | Explore more options |
| Subvocal | $499 early bird (US only) | None | Silent speech to text | Reserve on the official site |
The Ray-Ban Meta (Gen 2) remains the most polished pick in India, with a 12MP camera, 3K video and Meta AI. You can see the latest Ray-Ban Meta Gen 2 price here.
On a tighter budget, the Pebble Seeq brings a camera, open-ear audio and live translation under the cost of most phones. Our Pebble Seeq AI smart glasses launch coverage covers its full specs.
If a camera on your face bothers you, Subvocal is not the only camera-free idea around. XGIMI took a similar privacy-first route with display-based glasses, explained in our story on MemoMind One camera-free AI glasses.
For a homegrown option with Gemini and UPI built in, read our breakdown of B by Lenskart smart glasses.
FAQs on Subvocal and Silent Speech Technology
Does Subvocal read my thoughts? No. It only transcribes words you deliberately mouth through it. The company says there is nothing to transcribe if you are not actively trying to say something.
Does Subvocal use a camera? No. It relies on radar-based, non-contact sensing of your speech muscles, so it does not watch your lips.
How much do Subvocal glasses cost? The early bird price is $499 in the US, reserved with a $199 refundable deposit. Prescription lenses cost extra.
When will Subvocal ship? The company estimates the first shipments in the first quarter of 2027.
Is Subvocal available in India? Not yet. Reservations are open only in the US, and there is no announced India launch.
What is silent speech technology used for? Private messaging in public places, hands-free note taking, talking to AI assistants in quiet offices, and potentially helping people who find speaking or typing difficult.
The Bigger Picture for Silent Speech Interfaces
Keyboards have barely changed in decades, and voice assistants never solved the awkwardness of talking to a gadget in public. Silent speech technology sits right in that gap, and Subvocal is one of the first startups to wrap it into something that looks like regular eyewear.
Whether it becomes the next everyday input method depends on numbers the company has not shared yet: real accuracy, battery life and how well it understands accents.
The first shipments in early 2027 will answer those questions. Until then, it is easily one of the most interesting wearables to keep an eye on.