Adaptive Speech Controls:  Inclusive Design for Every Voice 

Written By: CEO Alex Dunn & CLO Julia Franklin 

Introduction 

The voice assistant revolution has changed the way we interact with technology. From asking ‘Alexa’ to play our favorite tunes to telling Google Home to turn off the lights, voice commands are becoming second nature. Yet, for many, voice-to-text technology is more than a convenience—it’s a vital tool for employment, daily living, and yes, even play.  

Speech-to-text software transforms spoken words into electronic written text, sometimes known as dictation. Some programs also provide voice-controlled navigation of a device. But too often, these solutions come with a hefty price tag and a one-size-fits-all approach.   

Our solution to a more inclusive approach? Meet Cephable: the versatile software that grows with you. Built on custom AI that learns from diverse speech patterns, Cephable offers a suite of tools that evolve with your unique needs.  We are pioneering a solution that allows users to interact with their tech in ways that work best for them. Individuals can use our software on their computers at no cost to them for non-commercial use.  Cephable also integrates our custom language models with enterprise software, so our controls are native to their platforms on mobile devices, gaming consoles, media boxes, or anywhere users could benefit from new inputs. 

An Inclusive Approach 

Your voice is much like your fingerprint—unique and unmistakably yours. Cephable honors this by providing adaptive voice controls that are responsive to your specific speech patterns. It’s not just about understanding you; it’s about providing a tailored, intuitive experience that grows and adjusts as you do. 

How did we get here?  

Our very first approach to using voice to control apps and games was done before Cephable was even a clearly defined idea. We started by hacking an Amazon smart speaker with an Arduino device as a sort of “voice-based keyboard” device called “Project SuaveKeys”. While extremely crude in its design, and built as part of a hackathon, it showed a path forward and a clear need for voice controls and speech recognition that worked better for folks with different speech patterns. From there, we set out to approach speech recognition in a new way – a way that adapts to each user and their context and one where the user had more control over how it worked. With a mix of transformer and graph-neural-net approaches, we can create extremely fast, offline, and private voice controls that adapt to you. 

The creation of Cephable’s language models is as much an art as it is a science. We’ve compiled vast arrays of vocal variations, teaching our AI to appreciate the nuances in every user’s voice. This dedication means Cephable’s technology is robust enough for a wide range of vocal tones and styles. 

Where are we now?  

With Cephable, you choose how you interact in your digital environment. Full dictation mode allows you to transcribe large chunks of text effortlessly. You can navigate your computer with intuitive voice commands for common tasks, like moving your cursor or clicking the mouse. A simple word like, “type” cues Cephable to switch into dictating your words again. Voice controls and dictation can be used alongside any app or program that you use on your computer…word processing, browser navigation, email, media, games…you name it!  

Need some examples? Every key on the keyboard has an out-of-the-box voice command, (e.g. tab, space, question mark, etc.). Or you can string together multiple keystrokes into one command, (Command+C = “copy”)

But, what happens if your speech accuracy changes over time? No problem. Customize commands to fit your speech.

Hard to say the /s/ sound? Simply choose voice controls to navigate your computer without using any words with /s/. Are multisyllabic words inconsistent? Change a command like, “period” to “dot”. Plus, the Optimistic Mode setting intelligently predicts your needs, activating the nearest command even if your speech isn’t perfect.  In this mode, Cephable will focus only on the voice controls in your selected profile. Optimistic mode is for users who are not in active dictation mode but are trying to trigger their voice controls faster, regardless of their speech accuracy.  

Where are we going?  

We’re just at the beginning of our story in revolutionizing voice technology. We’re adding more languages that we support with our adaptive controls, faster setup and personalization, more accurate dictation in complex scenarios, and also looking at ideas to turn our adaptive voice controls into easier ways for folks to communicate with others. Cephable is set to embrace even more of your world, keeping the focus on what matters—your voice. 

Success Stories

The true measure of Cephable’s success lies in the stories of our users:  

• One user didn’t let the “r” sound stand in his way! He created a custom profile for his Hogwarts Legacy game without any “r” sounds in his voice commands. 
• Another Cephable user with muscle spasms in his neck uses voice controls to manage his mouse clicks. With Optimistic Mode, he can have this helpful companion to fill in during his voice’s less reliable moments.   
• A non-speaking user who can vocalize a variety of vowel sounds was able to play his first browser game by mapping vowels/syllables to simple keyboard game controls. 
• Consider the creative freedom another user gained. With the use of just one arm, he blended single manual key presses with Cephable’s voice controls to open options for new programs in his digital landscape. 

Cephable In Action

Curious to see how this all works? Watch our users demonstrate voice controls and other inputs in their dance with digital commands in our recent showcase video.  

Inspired? Try Cephable for yourself and join our community that is shaping the future of voice-controlled technology.