Mozilla Common Voice
How do you make machines sound like humans? You let real people share their voices on a unified platform and create an open-source dataset with hundreds of thousands of voice samples.
When Mozilla decided to launch Common Voice, an initiative to improve voice recognition technologies, their main challenge was developing a dynamic and highly intuitive website that would make voice donation and validation a breeze.
You can donate your unique voice and help application programmers craft the next amazing real-time translators and voice-enabled digital assistants. All you have to do is read a sentence aloud and submit the clip for validation.
Our role in the project
We had a blast working with Mozilla on the Common Voice project. Their constant strive for innovation and our passion for all things digital have brought voice recognition developers a one-of-a-kind tool and narrowed the gap between humans and machines even further.
Our developers worked on coding various parts of their multilingual website, and make it accessible to everyone from any device and browser. We used React to build the core features and functionality that capture and store users’ voices with just a few clicks.
In the past, developers had to use expensive and proprietary voice data to build this kind of apps, but with the arrival of Common Voice, they now have an ocean of tones and accents to tap into.
Backed by a sleek design and outstanding ease of use, Common Voice has grown into a huge success story for which the sky’s the limit. The latest datasets contain over 13,000 recorded hours in 76 languages, including demographic metadata like age, sex, and accent that can help train the accuracy of speech recognition engines.