These smart glasses promise to turn your silent mouth movements into text
Reported by Root-Nation.com, Yahoo Tech, The Register and 1 more outlets4 sources ↓
Startup Subvocal is developing glasses that use radio signals to convert mouthed words into text. The Register reports preorders are open for $199 upfront, with another $300 due at shipping.
Key points
- Subvocal says its glasses can decode silent speech at more than 200 words per minute
- A $199 reservation is open; the early-bird price is $499
- The company says devices are expected to ship in Q1 2027
- Tech Times reports a $699 retail target and a $1.3 million pre-seed round
- Independent research has not demonstrated open-vocabulary RF silent speech at the company's claimed scale
Sources
- Root-Nation.comRead ↗A computer that understands without words: Subvocal is developing glasses for silent text input
- Yahoo TechRead ↗Subvocal’s glasses promise to let you talk to your computer in complete silence
- The RegisterRead ↗These smart glasses promise to turn your silent mouth movements into text
- Tech TimesRead ↗Subvocal Company Goes Viral: Smart Glasses Claim 200 WPM, Peer Review Caps at 50 Words
Details
The glasses use low-powered radio waves directed at the wearer’s face to detect movements involved in speech and turn them into text. The Register reports that the current prototype is not the final design and that release timing is not firm beyond the Q1 2027 estimate on the preorder page.
The Register says buyers can reserve a pair for $199, with another $300 due at shipping, for a $499 early-bird price. Tech Times reports that Subvocal lists a $699 retail target and has raised a $1.3 million pre-seed round backed by Y Combinator and Afore Capital.
Subvocal claims the system can decode open-ended speech at more than 200 words per minute. Tech Times notes that published research on a related radar method tested a 50-word vocabulary, with reported accuracy of 99.17% in the same training session and 88.87% across sessions; it says peer-reviewed research has not demonstrated open-vocabulary recognition at Subvocal’s claimed scale. The company has not provided an accuracy rate for its product.
Written from coverage as of
Images
- The Register
- Tech Times