The Subvocal Company unveils smart glasses for silent speech text input Other

Hardware startup The Subvocal Company introduced smart glasses that decode silent facial articulations into text in real time without vocal sound or external cameras.

Founded by Georgia Tech students and backed by Y Combinator and Afore Capital, the team raised a $1.3 million pre-seed round to commercialize hands-free, silent prompting for AI agents.

What it is

• Radar and RF articulatory sensing. Miniature antennas embedded in standard-looking frames emit low-power radio waves through the skin, capturing subtle micro-movements of facial speech muscles and the tongue.
• AI speech translation. An onboard neural model maps acoustic-free muscular articulatory patterns into digital text at speeds exceeding 200 words per minute.
• 90% transcription accuracy. Subvocal claims approximately 90% accuracy, surpassing optical lip-reading models while approaching standard spoken voice dictation.
• Privacy without cameras or microphones. Because the sensors capture internal muscular movements rather than external video or audio, the system preserves user privacy in shared office and transit environments.
• Pricing and reservations. The company opened pre-orders at an early-bird price of $499 (standard retail $699), targeting initial production shipments in early 2027.

Device specifications

Sources

• The Subvocal Company Pre-order
• Aarush Yadav on X
• Y Combinator Startup Directory