When hands speak, Terp 360's AI listens, then translates

At just 25, Elly Savatia is changing how technology communicates with the deaf community. A tech optimist and social entrepreneur, Savatia is the founder and CEO of Signvrse, an assistive technology startup using artificial intelligence to make sign language more accessible.

The idea for Signvrse was born in 2020 during a nationwide robotics training tour. After representing Kenya at an international competition, Savatia and his team visited schools across the country. Repeatedly, they noticed the same problem: deaf students were struggling to keep up because a single sign language interpreter was often expected to serve hundreds of learners.

The gap was impossible to ignore.

Savatia asked a simple but powerful question: what if sign language interpretation could live inside a device?

That question became Signvrse.

Initially, the team experimented with AI that could recognize sign language. But after two years, they realized the technology wasn't ready for real-world use. Instead of forcing it, they listened to the deaf community—and what they learned changed everything.

Deaf people consume information visually. Rather than trying to interpret signs, Signvrse shifted to generating sign language, delivering information in the most natural way for deaf users.

The result is Terp 360, Signvrse's flagship platform. Often described as a visual Google Translate for sign language, Terp 360 converts speech or text into a 3D digital avatar that signs information in real time. The goal isn't just access—it’s dignity, independence, and equality.

"Communication is the foundation of life," Savatia says. "Without it, access to education, healthcare, and jobs becomes a struggle."

According to the World Health Organization, an estimated 1.3 billion people—16% of the world's population—live with a significant disability. Disability is a natural part of the human experience, yet people with disabilities continue to face deep inequities—stigma, discrimination, poverty, limited access to education and employment, and systemic barriers. Many experience poorer health outcomes, have twice the risk of conditions such as depression, diabetes, and stroke, face transportation barriers, and may die up to 20 years earlier than those without disabilities.

For the deaf community in particular, sign language interpreters are often scarce or expensive, leaving many excluded from vital services. Terp 360's AI-powered system allows sign language to be embedded directly into websites, classrooms, hospitals, and workplaces, making access constant, private, and independent. For many deaf users, this technology is not just a convenience—it is life-changing, enabling them to engage in education, access healthcare, participate in public life, and pursue jobs on equal terms.

Powered by AI, Terp 360 uses natural language processing, machine learning, motion capture, and 3D animation to translate spoken or written language into accurate, expressive sign language. Continuous testing with deaf users has helped the platform reach over 85% accuracy, with steady improvements toward near-human realism.

"We're not replacing human interpreters," Savatia explains. "We're filling the gaps, especially where interpreters aren't available."

Over the past year, Signvrse has refined its products, translated digital content, and launched Terp4Web, enabling organizations to integrate sign language directly into their websites. Clients already include the Kenya National Innovation Agency.

Kenya's policies support this push. Sign language is an official national language, and laws now require that five percent of jobs be reserved for people with disabilities, signaling a growing commitment to inclusion.

Building advanced AI in Africa comes with challenges, from data collection to infrastructure. Yet Signvrse has gained global recognition, winning awards from Microsoft, the Royal Academy of Engineering, and the Google GNI Accelerator.

At the core of the company is co-creation. More than 30 deaf individuals actively contribute to Signvrse's development, ensuring the technology reflects real-life experiences.

Looking ahead, Signvrse plans to expand across Kenya, translating both English and Swahili into Kenyan Sign Language, with ambitions for multilingual growth and building one of the world's most comprehensive sign language datasets.

"Inclusion can't be an afterthought," says Elly Savatia. "If we don't design for accessibility from the start, we leave millions behind.”

Through Signvrse, Savatia is proving that when hands speak and AI listens, technology becomes truly human. AI can be a powerful tool to break down barriers for the deaf community, providing equal access to information, education, and opportunities.

CGTN Africa reporter Nyawira Mwangi and photographer Gabriel Rotich also contributed to this report.