Follow our journey building the future of Arabic clinical documentation.
We are proud to announce Sada — a first-of-its-kind full-stack platform purpose-built for constructing high-quality Arabic medical speech datasets. Sada is the engine behind Medad's data collection pipeline, enabling clinicians to record, transcribe, validate, and annotate real-world doctor-patient conversations in Gulf Arabic.
Live microphone recording with real-time transcription, plus batch upload of audio files (MP3, WAV, M4A, WebM).
GPU-accelerated Whisper transcription, automatic speaker diarization, language detection, and AI-generated summaries with key points.
100% on-premise deployment. All patient audio and transcriptions stay on local GPU servers — zero data leaves the hospital network.
Sada supports full-text search across transcriptions, folder-based organization, multi-format export (TXT, JSON, SRT), and secure shareable links — all managed through an intuitive dashboard. The platform is engineered with AI-assisted development and designed to scale across multiple hospitals and research institutions.
Collaborate With Us
Hospitals, universities, and researchers — help us build the largest Arabic clinical speech corpus.
Clinical data collection is now underway across participating hospitals in Muscat. Our team is actively recording real-world doctor-patient conversations in Gulf Arabic, building the foundation for Medad's speech recognition and clinical note generation models.
Initial model training has kicked off using collected speech data. We are fine-tuning Arabic ASR models and local LLMs on Omani medical terminology, Gulf dialect patterns, and clinical note structures — all processed on our on-premise GPU infrastructure.
A multidisciplinary team of 11 clinicians, researchers, and engineers from across Oman's top medical institutions has come together to build Medad — the first Arabic ambient clinical documentation system.
Institutional Review Board application submitted for ethical approval of clinical data collection across participating hospitals in Muscat. A critical step toward starting real-world recording sessions.
Evaluating on-premise GPU servers and local LLM deployment strategies to ensure 100% data sovereignty for Omani patient information. Testing various hardware configurations for optimal performance.
Initial project concept developed and research methodology designed. Identifying key collaborators across Omani medical institutions and defining the technical architecture for the system.
We are building the largest Arabic clinical speech corpus and welcome partnerships with hospitals, universities, and research institutions across the Arab world.
Whether you can contribute clinical recordings, linguistic expertise, or engineering resources — we want to hear from you.