Google details how Pixel 11 will translate ASL into text in real time

Google is rolling out a new accessibility feature for its Pixel 11 lineup that translates American Sign Language into on-screen text in real time. The company says the capability is built around a multilingual sign language-to-text, or SL2T, translation model that interprets signing motions and expressions and outputs written English sentences as the conversation happens.

In the announcement, Google describes the underlying approach as recognizing multiple visual indicates at once, including hand shapes, facial expressions, and body movements. It frames the feature as an ASL-first tool intended to let people communicate in their native language and then read the resulting English text immediately, without switching to speech or typing as a primary mode.

The translation feature is positioned as usable in more than one way on the Pixel 11. Google says the system supports both one-handed and two-handed signing, and it can be used with Live Transcribe, as well as through Gboard. Live Transcribe is Google’s on-device or device-assisted transcription tool that turns spoken input into text, and Gboard is the company’s keyboard app that provides text entry options and tools across Android.

Google says that to activate sign-to-text on a Pixel 11 device, users add sign-to-text to the Gboard toolbar. From there, the signer uses the front camera while signing in ASL. The company states that the translations appear on the device’s outer display, indicating the feature is designed to work in a viewable, face-forward setup during real-time interaction.

Google also says the technology was developed in partnership with the Deaf community. That partnership is presented as a key element of the product design, with the company saying the feature was created to reflect real-world signing needs rather than treating sign input as an afterthought to speech or typing.

The company’s post emphasizes that the new capability is powered by the SL2T model, which it describes as multilingual and specifically focused on translating sign language to text. A multilingual model, in this context, is designed to handle inputs across languages rather than restricting translation to a single sign language, although Google’s announcement is centered on ASL for Pixel 11 usage.

For Google, accessibility features like real-time translation can serve both a product mission and a platform strategy. Pixel devices are an entry point into Google’s broader ecosystem, including Gboard and transcription tools like Live Transcribe. By bringing sign-to-text into that ecosystem, the company is effectively linking hardware input (camera-based sign capture) to software output (English text in commonly used interfaces).

Even with the details provided, Google does not disclose performance measurements in the announcement, such as error rates, latency targets, or how accuracy varies by signer, lighting, background conditions, or sign speed. The company also does not specify whether the translations are processed fully on-device or through any network-assisted approach, nor does it describe any training data boundaries beyond the stated partnership with the Deaf community.

The next step for users and observers will be practical testing of how quickly and reliably translations appear on the outer display, and whether the feature behaves consistently across different signing styles and environments. Google says Pixel 11 is available on shelves now, and that prospective users can check the Google Store for more information and product updates as the rollout continues.