Key Notes
- Subvocal says its prototypes decode deliberate silent speech at more than 200 words per minute, although the launch announcement does not provide a numerical accuracy benchmark.
- The company describes RF sensing of tiny facial movements, rather than thought reading or camera-based lip reading.
- The $499 early-bird price requires a $199 refundable deposit, with US-only paid preorders and Q1 2027 deliveries subject to FCC approval.
Dictating a long message is convenient until everyone on the train can hear it. The Subvocal Company is building glasses that aim to remove that problem: silently articulate the words, and have them appear as text on a computer. The San Francisco startup says its prototypes can decode more than 200 words per minute, with reservations now open at a $499 early-bird price.
That is a claim about technology still being developed, rather than the performance of a finished product on sale. Founders Aarush Yadav and Arnav Kumar have shared a working demonstration, while the company’s website presents the sleeker design it wants to ship. The first deliveries are estimated for the first quarter of 2027.
Silent Speech, Not Mind Reading
Subvocal’s input comes from deliberate articulation. You make small movements associated with saying words, without speaking aloud, and the system interprets them. Its product FAQ explicitly rules out reading arbitrary or involuntary thoughts. Thinking about a sentence and intentionally mouthing it are different actions.
In their launch announcement, the founders describe RF sensing: low-power radio signals track tiny facial muscle movements, and a trained model translates the resulting signals into text. The approach does not require a camera to watch the wearer’s lips. The movements can be subtle, although the company stops short of saying nobody nearby could notice them.
The practical attraction is easy to understand. A headset microphone can make ordinary dictation more convenient, but it cannot stop someone sitting next to you from hearing a private message. A silent input device could let someone draft an email, write notes or give an AI assistant a detailed instruction without broadcasting the words across an office.
The 200-Words-Per-Minute Claim Needs More Testing
The headline speed deserves some restraint. The founders report open-ended decoding at 200+ words per minute, but describe the error rate qualitatively rather than publishing a numerical accuracy benchmark in their launch post. A demo can show that an interface works; it cannot establish how reliably a production device will handle different wearers, unfamiliar vocabulary or a full day of use.
Subvocal acknowledges that its prototypes make mistakes and says the product is being designed around context and easy corrections. The founders also plan to collect a much larger dataset to develop a model that needs only brief personal calibration. Broad performance across users remains a development goal.
Correction speed will be particularly important for anything sold as an alternative to a keyboard. A fast stream of text is less useful if the wearer repeatedly has to stop, reach for a phone or laptop and repair a name or command. Comfort, calibration and dependable input in existing apps will matter alongside the words-per-minute figure.
It is a different bet from adding a spoken assistant to a pair of glasses. Our coverage of Qwen glasses examined that more familiar approach. Subvocal instead addresses a limitation in voice interfaces: people may want the speed of dictation without making their conversation public. Whether silent articulation becomes a habit will depend on how little effort it takes.
What the $499 Reservation Actually Includes
The company’s preorder page lists a $499 total early-bird price, discounted from $699. Buyers pay $199 now and the remaining $300 later. The deposit is refundable until full payment is made; it is not the full purchase price.
Paid preorders are currently limited to the United States. Taxes, shipping and optional prescription lenses cost extra, and delivery is subject to FCC approval. The stated Q1 2027 shipment window is an estimate. For now, Subvocal is offering an early reservation for an emerging input device, with a working prototype behind it and substantial work still ahead of a finished pair of glasses.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
