What happened
Falcon-ASR, a speech-to-text system with 1.6 billion parameters, was unveiled on October 7, 2026. It transcribes Arabic audio and places special emphasis on the Emirati dialect. At launch, visitors could experiment with its transcription through a Hugging Face Demo Space, while API access and native apps were slated for later.
Why it matters
The demo Space lets people evaluate transcription quality and also offers word-level timestamps that map every recognized word to its spot in the recording. Programmatic access through an API and standalone apps had not been released at launch.
Read the original at Hugging Face