Egypt Launches BelMasry, a Free State-Built AI Platform for Egyptian Colloquial Arabic and 50 Foreign Languages

Egypt has formally launched BelMasry, a free, state-built artificial intelligence platform designed to process Egyptian colloquial Arabic and Modern Standard Arabic, and to translate between them and 50 foreign languages. Built by the Communications Ministry’s Applied Innovation Center, the pilot platform combines speech-to-text, machine translation and text-to-speech engines and is available to individuals, institutions and developers at belmasry.aic.gov.eg.

“BelMasry is a national system capable of understanding Egyptian colloquial Arabic,” Communications Minister Raafat Hindi said when the technology was previewed in February, framing the platform as the country’s first sovereign AI system of its kind.

The ministry describes BelMasry’s architecture as three integrated engines. The speech-to-text module transcribes spoken Egyptian Arabic across regional dialects and Modern Standard Arabic, distinguishing between speakers in multi-party recordings and filtering background noise. The machine-translation engine supports exchanges between Arabic and 50 foreign languages. The text-to-speech component renders written Arabic into natural-sounding speech, with use cases targeted at virtual assistants, digital education, media production and voice-enabled government services.

• Developer: Applied Innovation Center (established 2019)
• Access: Free at belmasry.aic.gov.eg for individuals, institutions and developers
• Core features: Speech-to-text (dialect-aware), machine translation (Arabic ↔ 50 languages), text-to-speech

The project aims to close a measurable language gap. Arabic represents less than 1% of global large language model training data, UNESCO figures show, and the portion covering colloquial dialects is even smaller. Egyptian Arabic is spoken by well over 100 million people and is the most widely understood Arabic dialect across the region—yet it remains a low-resource language for machine learning. The ministry says BelMasry was trained on Egyptian and Arabic linguistic datasets to better recognise local expressions, cultural references and emotional nuance that global systems often miss.

BelMasry joins a wider national portfolio developed by the Applied Innovation Center. The platform was previewed at an AI summit in Cairo alongside Karnak, Egypt’s national large language model; Sia, an AI tutor for secondary students; Tarjuman, a translation app; and locally trained diagnostic tools for diabetic retinopathy, macular edema and breast cancer. The government is also working with the UN Development Programme to share these tools through the AI-Share initiative, making the underlying approach exportable to other Arab and African states.

Egypt’s private sector has pursued parallel efforts: Nanovate raised $2 million to build Arabic-native voice and chat agents covering 22 dialects; Intella developed an Egyptian-dialect voice ordering system for Jumia Egypt before expanding into Swahili; and Alexandria-based TokenAI has released open-weight models for colloquial Egyptian, Levantine, Gulf and North African variants. What sets BelMasry apart, officials emphasise, is its sovereign, state-built and freely available model—more akin to Nigeria’s open-sourced national model than to commercial offerings.

Performance and adoption remain the key tests. The ministry has not published accuracy benchmarks for speech recognition across Cairene, Sa’idi, Alexandrian and Delta variants—metrics that experts say are necessary to evaluate real-world utility. BelMasry’s launch sits inside Egypt’s National AI Strategy 2025–2030, which targets 30,000 trained AI specialists and an AI contribution of 7.7% to GDP by 2030. Egypt ranked first in Africa and 51st of 195 countries on the 2025 Government AI Readiness Index compiled by Oxford Insights, climbing 14 places year-on-year.