Subvocal smart glasses turn silent mouth movements into text

A startup named "Subvocal" is developing smart glasses designed to translate facial and mouth muscle movements into computer text and commands. The technology allows users to type messages or interact with AI assistants without audible speech or a physical keyboard, by detecting the subtle motions associated with forming words silently.

Pre-order details and shipping timeline

The company has opened pre-orders in the United States at an early-bird price of $499. It plans to offer prescription lenses for an additional cost and targets shipping commencement during the first quarter of 2027. In an official announcement, Subvocal stated that its prototypes can convert speech movements into text at speeds exceeding 200 words per minute, though the device remains in the development phase prior to mass production.

How the silent input works

The concept relies on the fact that forming words requires moving muscles responsible for articulation, even when no sound is intentionally produced. Subvocal aims to capture these movements and convert them into digital inputs understandable by computers. According to an announcement published via the "Y Combinator" platform, the glasses use radio frequency sensing to detect small facial changes occurring when a user shapes words silently.

The system sends low-power radio waves across the face and measures alterations caused by muscle movement. Machine learning models then transform these signals into text. Users must still deliberately form words using articulation movements, but without vocalizing them. The company notes that these gestures can be either obvious or so limited that they are difficult for someone sitting opposite to notice.

Targeted use cases and calibration

The glasses aim to facilitate text entry in situations where voice dictation is impractical, such as shared offices, cafes, and public transport, where speaking to a computer might disturb others or reveal message content. Proposed applications include writing emails, notes, and messages; entering instructions into AI tools; and coding or issuing commands. The core function is converting silent speech into digital input, while the execution of requests depends on the software receiving those inputs rather than the muscle sensor alone.

Currently, the company is working on personalizing the model for new users through individual calibration data, a step where the system learns the specific characteristics of the user's signals.

Competitive landscape and Apple acquisition

SubVocal enters a market that has recently begun to take shape, facing intense competition. This year saw Apple acquire "Q.ai," a company working on technologies related to understanding speech and fine facial movements. Reuters reported the deal’s value at $1.6 billion. Approximately 100 employees joined Apple, including CEO Aviad Maizels and co-founders Yonatan Wexler and Avi Bar-Lea.

Maizels previously founded the 3D sensing company "PrimeSense," which Apple acquired in 2013. Patent documents linked to Q.ai describe capturing silent speech using optical sensors that monitor micro-movements of facial skin. One patent, titled "Silent Speech Detection," outlines a device directing light toward a facial area, such as the cheek, and analyzing reflected light changes caused by skin movement to generate speech outputs matching the user's formed words.

Xiaomi’s wireless AR prototype

At MWC 2023 in Barcelona, Xiaomi unveiled a new prototype for its Wireless AR Glass Discovery Edition. This third-generation model follows initial versions released in 2021 and August 2022. The glasses feature two MicroLED screens with brightness reaching 1,200 nits and a cone-based system to reconstruct images before the user's eyes. Display resolution reaches 58 pixels per degree of vision, closely approaching the human retina’s estimated 60 pixels per degree.

The prototype includes electrochromic lenses that adjust visual clarity based on ambient lighting and can darken completely to switch to VR mode. Connectivity is streamlined via NFC, allowing pairing by tapping the back of a phone against the glasses' arm. Xiaomi highlighted compatibility with devices supporting the Qualcomm Spaces platform, such as the Xiaomi 13 and OnePlus 11. The company noted its proprietary connection technology achieves data transmission latency as low as 50 milliseconds.

Powered by the Snapdragon XR 2 Gen 1 processor and supporting the Snapdragon Spaces XR platform—similar to Meta’s Quest Pro—the glasses utilize Mi Share for streaming content like YouTube and TikTok videos from phones. A developed AR Gesture Control system enables navigation via hand and thumb movements for scrolling, as well as controlling smart home devices like lights. Xiaomi confirmed the model remains experimental, with no precise market launch date set.

Samsung’s auto-focus lens patent

Samsung revealed advanced technology for smart glasses with automatic corrective lenses, detailed in a recent patent filed with the United States Patent and Trademark Office. Announced during the Unpacked 2025 event alongside a roadmap featuring foldable phones and mixed reality headsets, the innovation addresses users requiring prescription eyewear.

The patent describes a sophisticated optical system allowing automatic lens adjustment to improve vision. This mechanism employs gears and precision motors to reposition lenses for clearer images and expanded field of view. While this represents a significant leap for wearable tech—particularly for those wanting VR or AR experiences without extra glasses—commercial viability faces challenges, as patent registration does not guarantee product release.

Regarding Project Moohan, leaks suggest it will compete with Apple’s Vision Pro, offering a more refined design and innovative features. Samsung collaborates with Google and Qualcomm for this headset; Google has introduced its Android XR operating system, and Qualcomm promised a revolutionary chip tailored for the device. Additionally, Samsung integrated the HTC VIVE glasses development team into its own, strengthening Project Moohan’s position in the mixed reality market.

The smart glasses sector has seen substantial growth since 2023, with major tech firms and startups striving to mature the technology into a commercially viable product capable of delivering distinct user experiences.