• News
  • Middle East

HUMAIN Launches Arabic-First Speech AI Platform

Platform combines ASR, TTS, voice agents, and an open-weight model.

11/19/2026
•Ghita Khalfaoui
Back to News

HUMAIN has launched HUMAIN Voice, an Arabic-first speech AI platform. The platform combines automatic speech recognition, text-to-speech, and voice agents. HUMAIN, a PIF company, notes that more than 450 million people speak Arabic worldwide.


Open-weight model

The company also released HUMAIN-Voice-Lite-80M. This open-weight ASR model has 80 million parameters and uses an Apache 2.0 license. It runs on both CPU and GPU infrastructure for edge and on-device use.

Dialect and code-switching support

The platform supports Saudi dialects such as Najdi and Hijazi. It also covers Gulf dialects, English, and natural Arabic-English code-switching. The design treats dialect, accent, and context as core speech features.

Voice agents for enterprises

Organizations can use HUMAIN Voice agents to build full-duplex speech-to-speech systems. These agents listen, understand, and reply in the caller's own dialect with low latency. HUMAIN's ASR and TTS capabilities power a wider range of products and applications.


HUMAIN will release benchmark results from the Open Universal Arabic ASR Leaderboard. It will also publish a blog on the model's recognition across Arabic, English, and code-switching. Developers and enterprises can access the open-weight model under the Apache 2.0 license.