Subvocal develops smart glasses for silent text transcription

San Francisco-based startup Subvocal has introduced smart glasses designed to convert silent lip movements into text. Using low-power radio waves that pass through facial tissue, the device's sensors detect muscle movements associated with articulation. An AI model then decodes these signals into text at a reported speed of over 200 words per minute.

Founders Aarush Yadav and Arnav Kumar developed the technology, which aims to allow users to interact with computers, write messages, or prompt AI assistants without using a keyboard or speaking aloud. The company emphasizes that the device does not read thoughts, but only transcribes words that the user chooses to articulate.

The project, which emerged from a Y Combinator program, has secured 1.3 million dollars in funding. While the current hardware is still in development, pre-orders have opened in the United States. Early bird pricing is set at 499 dollars, with a standard retail price of 699 dollars and estimated deliveries starting in March 2027.