Voice AI

Voice AI

Inspiration

Voice AI - A Voice for Every Child to learn, read and speak offline in 70+ languages. My journey creating an AI Voice App for Children in rural Africa. In the green corners of rural Africa, where farming is a tradition and classrooms are scarce, textbooks and learning are often a luxury. I saw more than just a gap in education—I saw unheard voices. Some of these voices belonged to children with visual and speech impairments. Most simply lived too far from schools they could never afford. But all of them shared one thing in common: the right to learn and share their visions with a voice. That belief drove me to build something simple - not with walls or chalkboards, but with AI and the power of voices that speaks more than 70 languages.

💡 The Vision I asked myself: What if we could bring learning to every child, without needing expensive schools, textbooks, electronics or even the internet? So I set out to create Voice AI - a tool that could teach children anytime, anywhere, using generated AI voices of textbook, handwriting and texts. But we didn’t want just another app. We wanted a simple experience that felt easy, inclusive, and deeply free - even for a child who had never used any kind of technology before, to have the freedom to learn without classrooms, internet, mobile phones, expensive technologies or electricity.

🚸 No Child Left Behind. This project wasn’t about technology for technology’s sake. It was about equity. About giving every child, regardless of disability or geography, a way to explore their potential through voice or speech. Because education shouldn’t be a privilege. It should be a birthright for those with or without a voice.

🔥 The Impact we have seen in just a week - not from the glow of a tablet, but from the joy of being heard and listening to audiobooks without limit. We have watched speech impaired children speak using AI voices with growing confidence, giggling as they answer questions or sing along to learning songs. And we have heard from families who now feel hope where there was silence.

Because when we empower children to learn with the tools they already have—voice, curiosity, and courage—we don’t just teach lessons. We ignite futures.

What it does

Voice AI generates answers to questions or learning topics and transforms and translates all generated texts, handwriting, inscriptions and textbooks into audiobooks and speeches in over 70 languages, allowing the audio to be downloaded, shared, accessed and played on any audio device without needing an internet connection or an expensive device like phones, anytime and anywhere.

How we built it

🤖 We harnessed the power of AWS services like Bedrock, Textract, Polly and Translate to create an AI natural language processing in more than 70 languages for text and voice generation, handwriting and text recognition, and translation for children anywhere to learn. For blind children, they could now read and learn with voices that teaches them every second and every day without restrictions or boundaries. For children with speech impairments, they could gently speak with clear voices, offering understanding instead of frustration. And for children with no access to internet, it works offline, voices generated can be downloaded and shared without the internet, making learning available in the remotest places.

Challenges we ran into

Picking the best technology partner was a challenge, coming from the African continent, some cloud companies block access to some of these technologies. We were rejected credits and services by Oracle and Google but found Amazon Web Services to be more helpful. We ran into a terrible bug building with Amazon Bedrock. We couldn't access the models with the bedrock API that had colon ":" like amazon.nova-micro-v1:0. We finally resolved this issues and got the app working.

Accomplishments that we're proud of

We successfully developed an MVP of our vision and shipped to Play Store after 2 weeks of tough learning curves and sleepless nights. AWS was very helpful during this process but there were problems even the support team couldn't solve that we had to figure out ourselves.

What we learned

We learnt how to build easily in Kiro and integrate AWS services in less than a week. We also discovered using Generative AI API are very expensive. During the short testing period of our app, we collaborated with teachers and parents in our community to test our app with their kids, who helped us improve on the app as they continue to enjoy their newly found learning style and voice as they learn.

What's next for Voice AI

🚀 The Journey Continues - this is just the beginning. Our mission is to continue improving the Voice AI app, integrating more languages like our local dialects, and Integrating more robust services that will make learning universally accessible to every child regardless of their social, religious or economic status. We also plan to make our solution available on IOS, Linux, MAC and Windows platforms.