Person using a laptop with a webcam in a dim room
Accessibility

Project GameFace Lets People Control a Computer With Only Their Face

7 min read|Updated November 2026
Share

At a glance

Project GameFace is a free, open source AI tool from Google that turns head movements and facial gestures into full cursor control, using nothing more than a standard webcam. Built with a quadriplegic gamer and now available for both Windows and Android, it replaces head tracking gear that can cost thousands of dollars with software anyone can download.

  • Launched at Google I/O in 2023 and extended to Android in 2024, with all code open sourced
  • Co-designed with Lance Carr, a quadriplegic streamer living with muscular dystrophy
  • Users can raise an eyebrow to click, open their mouth to move the cursor, and define their own gestures
  • Runs locally with Google MediaPipe machine learning models, so no cloud connection is required
  • Companies like playAbility have built inclusive software on top of its open source building blocks

Project GameFace is a hands free AI mouse from Google that lets people control a computer or Android phone using only head movements and facial gestures. It matters because it replaces expensive assistive hardware with free, open source software that runs on any ordinary webcam, giving millions of people with limited hand movement independent access to technology, games, work, and communication.

The idea for GameFace came from a livestream. Lance Carr, a quadriplegic video game streamer who lives with muscular dystrophy, a progressive disease that weakens muscles, was gaming one night in 2021 when his house caught fire. The blaze, which was broadcast live to viewers around the world, destroyed his home and all of his equipment, including the expensive head tracking mouse that made gaming possible for him. Engineers at Google saw the story and reached out, and the result was a collaboration to build something better and radically cheaper.

How a Webcam Becomes a Mouse

GameFace is built on Google's open source MediaPipe framework. Machine learning models track the user's face through a standard webcam and translate what they see into cursor actions. Moving your head steers the pointer across the screen. Raising an eyebrow can trigger a click. Opening your mouth can start or stop cursor movement. Because everything runs locally on the device, there is no cloud round trip, which keeps the tool responsive and private.

One design decision separates GameFace from rigid assistive tools. Users choose which facial expressions map to which actions, and a setting called gesture size controls how pronounced a movement must be to register. A person with very limited mobility can set small, subtle gestures, while someone with more range can use exaggerated ones. The tool adapts to the body rather than forcing the body to adapt to the tool.

From Gaming to Phones

Google first released Project GameFace at I/O in 2023 as a Windows tool aimed at gamers, and the code was open sourced from the start. In May 2024 the team went further and open sourced additional code so developers could build hands free control into Android apps. The Android version creates a virtual cursor driven by head movement and uses the Android accessibility service to support system level actions such as going home, going back, opening notifications, and switching apps.

Under the hood, the Android release uses MediaPipe's Face Landmarks Detection API, which produces 52 blendshape values representing individual facial gestures such as raising the left eyebrow or opening the mouth. Developers can set different thresholds on each expression, so the experience can be tuned precisely for each user. Google built this release with Incluzza, a social enterprise in India that supports people with disabilities, to test how the tool performs beyond gaming, in settings like typing messages to family or searching for jobs.

Why Free and Open Source Matters Here

Dedicated assistive technology has long carried a painful price premium. Eye tracking systems and specialized head mice can cost thousands of dollars, and insurance often refuses to cover them, particularly for conditions classified as progressive. When Lance Carr lost his equipment in the fire, replacing it was a serious financial blow, not a quick shopping trip.

GameFace inverts that model. The software costs nothing, runs on hardware most households already own, and lives on GitHub where anyone can inspect, extend, or fork it. Other companies have already built on top of it. playAbility, a studio making inclusive gaming software, uses GameFace building blocks such as the MediaPipe blendshapes in its products. Because the code is open, the work compounds instead of staying locked inside one company's roadmap.

The Bigger Picture for Accessibility

The World Health Organization estimates that more than one billion people live with some form of disability, and access to digital technology increasingly determines access to education, employment, and community. Facial control is not the right answer for everyone. Some conditions affect facial muscles as well as limbs, and gaze based systems remain essential for many users. But for people whose faces work well while their hands do not, GameFace offers something rare: full, dignified, independent control of a mainstream device at zero cost.

The name is playful, but the stakes are not. A person who can operate a computer alone can play games with friends, apply for jobs, order groceries, and send messages without waiting for a caregiver to be free. Google could have kept this as a research demo. Instead it shipped it twice, open sourced everything, and handed the building blocks to a community that is already running with them.

Common Questions

What is Project GameFace?

Project GameFace is a free and open source hands free mouse from Google that uses machine learning and a standard webcam to translate head movements and facial gestures into cursor control. It was launched for Windows at Google I/O 2023 and extended to Android in 2024.

Who inspired Project GameFace?

The project was co-designed with Lance Carr, a quadriplegic video game streamer who lives with muscular dystrophy. After a house fire in 2021 destroyed his expensive head tracking mouse during a live stream, Google engineers partnered with him to build a low cost replacement.

How does Project GameFace work?

It uses Google MediaPipe machine learning models to track facial expressions through a webcam. Head movement steers the cursor, and gestures such as raising an eyebrow or opening the mouth trigger clicks and other actions. Users can customize which expressions map to which actions and how pronounced each gesture needs to be.

Is Project GameFace available on Android?

Yes. In May 2024 Google open sourced code that lets developers build GameFace hands free control into Android apps. It creates a virtual cursor, supports system actions like home, back, and notifications through the Android accessibility service, and was developed with input from Incluzza, an Indian social enterprise supporting people with disabilities.

How much does Project GameFace cost?

Nothing. The code is fully open source and available on GitHub, and it runs on a standard webcam or phone camera. This contrasts with dedicated head tracking and eye tracking hardware that can cost thousands of dollars and is often not covered by insurance.

Sources: Google Developers Blog, Project GameFace launch and Android release posts (May 2023, May 2024); The Keyword, Google blog announcing Project GameFace (May 2023); google/project gameface on GitHub; The Verge coverage of Project GameFace (May 2023); World Health Organization disability fact sheet.