You Can Now Use Stephen Hawking's Speech Software for Free

The code that gave Stephen Hawking his voice is now open source. Here's how you can use it.
Stephen Hawking.
Ted S. Warren/AP

Software created by Intel was instrumental in giving Stephen Hawking a voice. Now, the company has released this same software under a free software license.

The development of the platform, called ACAT for "assistive context-aware toolkit," was detailed in a WIRED story earlier this year. It's a system that makes computers more accessible to people with disabilities. And now that the source code for this toolkit is open, it means you can build a system very similar to the one Professor Hawking uses to input text, send commands to applications, and communicate with the world.

So why aren't you? Well, there are a few things you should be aware of before you go ahead and download ACAT. For starters, it's PC-only. You will need a PC running at least Windows XP—and there are no plans to bring ACAT to Macs. Sorry, Apple users everywhere.

If you have a PC, though, the rest of the hardware requirements are pretty easy to meet. ACAT uses visual cues in the user's face to understand commands—in the case of Professor Hawking, who has ALS, it tracks the movements of his cheek muscle. To use it, your computer simply needs to have a webcam. However, for users who might want or need more from ACAT, there are possibilities for other types of input down the road.

"We have been busy building different sensors and trying this out with patients," says Intel principal engineer Lama Nachman, who leads the Anticipatory Computer Lab at the company. Nachman says this experimenting includes proximity sensors, accelerometer-based sensors, and Intel's RealSense 3D camera. So if you have any of those things, then ACAT could become vastly more interesting in the future.

Of course, ACAT isn't really meant for the average user to download and play with—at least, not yet.

"The goal of open sourcing this is to enable developers to create solutions in the assistive space with ease, and have them leverage what we have invested years of effort in," says Nachman. "Our vision is to enable any developer or researcher who can bring in value in sensing, UI, word prediction, context awareness, etc. to build on top of this, and not have to reinvent the wheel since it is a large effort to do this."

Nachman says Intel has sent some end users detailed instructions on setting up ACAT as well as how to use it, and the team is also working with some patients on testing it out. Soon, they'll be partnering with universities on experimenting with the system. It's clear that ACAT is for developers and the academic community, at least for now. While you certainly can go to Github and download it, Intel isn't catering to everyone here. However, early accounts make it seem as if it's fairly easy to install ACAT. You'll want to check out the user guide, of course, but then you're free to install, boot it up, and start letting your computer read your face and type for you.