ArabicSpeech

We welcome everyone working on Arabic speech - researchers, engineers, students, and practitioners, to use this space to share or link your models, datasets, benchmarks, tools, and demos. Our goal is a central, open repository spanning Modern Standard Arabic (MSA) and Arabic dialects (Egyptian, Levantine, Gulf, Maghrebi, etc.) across tasks such as ASR, TTS, and more. Publish your work here so the community can discover, reuse, compare, and build on it.

Mission

ArabicSpeech began as a volunteer and community effort to catalog Arabic speech resources and grew into a community hub for MSA and dialects. Over time, it expanded from curated lists to open datasets, benchmarks, and models, --- now hosted on Hugging Face, welcoming everyone to share, reuse, and build together.

How to share (quick)

Option 1 — Upload in the ArabicSpeech space

• Create a Model/Dataset/Space under ArabicSpeech.
• Add a license and tags (arabic, msa/dialect, task).
• Include a short README (what, usage, citation).

Option 2 — Add to an ArabicSpeech Collection

• Create a Collection under ArabicSpeech.
• Add links to HF repos or external resources (GitHub/Zenodo/papers) with a one-line note.
• Tag, publish, and share the Collection link.