In this long-waited release we introduce Speaker diarization, a feature that will elevate your transcription experience by a mile.

New Features

Speaker Detection (Beta)

Automatically identify and separate different speakers in your recordings. Speaker detection can be done via a new button in the transcript view.

⚠️ Please note that this process is computationally intensive and may take some time. We're eager to hear your feedback on this initial release!

Language-specific Voice Models

We've added specialised models for 🇫🇷 French, 🇯🇵 Japanese, 🇮🇱 Hebrew, and 🇳🇴 Norwegian, offering improved accuracy and lower error rates compared to standard multi-lingual versions.

Improvements

Enhanced Language Selection

Our redesigned language selector now features flags for each language, making it easier and faster to find the right language for your transcription.

Right-to-Left (RTL) languages support

Automatic RTL view switching for Hebrew and Arabic languages, enhancing readability and user experience.

Empty state

If for any reason QuickWhisper couldn't transcribe the source it will show a message clearly explaining the reason.