Intron Expands Sahara AI Speech Platform to 57 Languages

Nigerian artificial intelligence startup Intron has expanded its speech recognition platform, Sahara, to support 57 languages. The upgrade introduces 24 additional languages and strengthens the company’s ambition to build voice technology designed specifically for African communication patterns.

Sahara v2 now includes support for 23 African languages and more than 500 regional accents. Newly added languages include Hausa, Swahili, isiZulu, Yoruba, Kinyarwanda, Twi, Igbo, isiXhosa, Amharic, Luganda, Oromo, Pidgin, Shona, Wolof, and several others. The company selected these languages based on growing commercial demand from enterprise users.

The platform provides speech to text transcription, text to speech synthesis, and voice authentication tools. These services are designed for industries such as healthcare, financial services, legal systems, telecommunications, and government institutions.

Addressing Africa’s Voice First Technology Landscape

Africa has nearly 2,000 languages, and many exist mainly as spoken languages with limited written records. This linguistic diversity creates challenges for traditional digital systems that rely heavily on written text.

Voice recognition technology offers an alternative interface for interacting with digital platforms. By enabling speech based communication, AI systems can make services more accessible across regions where literacy levels and language diversity vary widely.

The global speech recognition market is expected to reach more than 81 billion dollars by 2032. African startups like Intron are developing localized infrastructure rather than adapting systems originally designed for Western languages.

Sahara v2 relies heavily on locally sourced voice data to capture the unique accents, dialects, and speech patterns found across the continent.

Performance Improvements and Benchmark Results

Intron reports strong performance improvements in its latest model. Internal benchmark testing using African voice datasets shows Sahara v2 outperforming several global speech recognition systems, including Gemini, GPT models, Whisper, ElevenLabs, and Azure.

According to the company, the system achieved up to 64 percent better recognition accuracy for African names, organizations, and geographic locations. The platform also demonstrated 35 percent improved handling of numbers and 20 percent stronger performance in noisy environments or multi speaker situations.

Cross industry testing showed around 25 percent higher accuracy across sectors such as healthcare, finance, legal services, and telecommunications.

Bilingual AI Models for Real Conversations

One of the most notable features of Sahara v2 is the introduction of a bilingual Swahili and English automatic speech recognition model. The system can handle code switching, which occurs when speakers alternate between languages within the same conversation.

Code switching is common across many African countries. For example, a doctor might ask questions in English while a patient responds in Swahili before switching languages again. Many traditional AI models struggle with this pattern.

Intron developed the bilingual model in partnership with Penda Health, a healthcare provider in Kenya. The technology supports clinical conversations and improves documentation accuracy in multilingual environments.

The company plans to introduce additional bilingual models for languages such as Yoruba, Hausa, Zulu, and Kinyarwanda in future releases.

Building African Voice Data Infrastructure

Intron was founded in 2020 by Tobi Olatunji and Olakunle Asekun. The company initially focused on clinical documentation tools before expanding into broader voice infrastructure for enterprise applications.

To train its models, Intron collected more than 14 million audio recordings representing over 50,000 hours of speech from more than 40,000 speakers across 30 African countries. Much of the early dataset had to be created from scratch because African medical speech data previously did not exist.

The company recruited contributors across multiple countries and compensated them for providing voice samples. Additional datasets supported by organizations such as the Gates Foundation, Lacuna Fund, and Google later expanded the training resources.

Expanding Use Cases Across Africa

Sahara is already used in several African countries, including Nigeria, Kenya, South Africa, Ghana, Rwanda, and Uganda. Enterprise clients include public institutions and private companies that rely on voice technology to process spoken information.

For example, the Ogun State Judiciary uses the platform for transcription services. In South Africa, the company Audere uses Sahara to transcribe WhatsApp voice messages across different regional accents.

Intron has also introduced a Hausa language text to speech model that can power multilingual voice assistants for call centers, healthcare systems, and financial services.

Offline AI Deployment for Low Connectivity Regions

Sahara primarily operates through cloud infrastructure. However, Intron now supports offline deployments through a partnership with Nvidia.

Organizations can run the models locally using Nvidia Jetson Edge devices. These systems cost about 250 dollars and allow multiple users to access the AI platform within local networks. This feature helps institutions operate in areas with limited internet connectivity.

The company also allows enterprise clients to choose whether their data is stored locally or in cloud environments to comply with national data protection regulations.

Future Expansion and Investment Plans

Alongside the platform upgrade, Intron plans to release its first Africa Voice AI Report in 2026. The report will analyze voice technology readiness, benchmarking standards, and data quality across the continent.

The startup also plans to raise around 3 million dollars in funding later this year. The investment will support expanded language coverage and further development of bilingual and industry specific models.

As demand for voice interfaces grows, companies like Intron are building critical infrastructure that enables African languages to function within global digital systems. The success of these platforms could significantly improve access to technology across multilingual communities.