Subvocal opens preorders for glasses that turn silent speech into text
The YC-backed startup claims more than 200 words per minute without a camera or audible voice.
Subvocal has raised $1.3 million from Y Combinator and Afore Capital and opened US reservations for glasses designed to convert silent articulation into text. The startup claims its prototype can exceed 200 words per minute without requiring the wearer to speak aloud or touch a keyboard.
The glasses use non-contact radio-frequency sensing to track tiny movements in speech muscles. An AI model interprets changes in the reflected signal and converts deliberate, almost invisible mouthing into text. The company says the system does not use a camera, lip-reading video or ambient audio. It is not presented as mind-reading: the wearer must intentionally form the words.
Subvocal pitches the device as an input method for messages, notes and AI assistants, especially in places where voice dictation would be awkward or disruptive. The early-bird price is $499, reserved with a refundable $199 deposit; the regular listed price is $699. Prescription lenses will cost extra, and the company expects initial shipments in the first quarter of 2027.
The product is still under development, and the claimed speed has not been independently verified. Accuracy across accents, names and long dictation sessions also remains an open question. If the system works as advertised, it could offer a private, high-speed interface for AI assistants without the social friction of talking to a computer.
