The Subvocal Company unveils wearable glasses for silent speech input

The most discussed startup of the weekend is The Subvocal Company.

The team introduced a wearable device for silent speech text input. It looks like ordinary glasses.

Inside the glasses are antennas that emit radio waves. The waves pass through the skin and capture movements of the muscles and articulatory apparatus when a person moves their lips and tongue without making sound. An AI model then converts the signal into text.

In the era of AI agents, dictating commands is often more convenient than typing prompts, while speaking out loud is not always practical. That helps explain why the device attracted so much attention: the announcement post on X has already received about 11.5 million views.

According to the founders, typing accuracy is about 90%. That is higher than typical lip-reading models but lower than ordinary voice text input, which usually reaches about 95–96%. The typing speed with this method exceeds 200 words per minute.

The founders, students at Georgia Tech, built the first prototype at a Stanford hackathon in February this year. They are now entering the YC 2026 fall batch, and their total pre-seed funding is about $1.3 million.

Preorders for the glasses are already available at subvocal.company.