Speech Note is an offline, AI-powered app able to transcribe your speech into text in a variety of different languages.
A reader got in touch to point me towards the app — thanks, David! — and given that it sounds pretty cool I figured I’d give it a spotlight on the site.
Speech Note use OpenAI’s Whisper and a stack of other open-source libraries, voice engines, and other doohickeys to perform its transliterative magic.
It supports Speech to Text (i.e you speak, it types), Text to Speech (i.e. you type, it speaks), and machine translation to translate text/speech from one language to another.
“Text and voice processing take place entirely offline, locally on your computer, without using a network connection. Your privacy is always respected. No data is sent to the Internet,” assures the application’s Flathub listing.
Those with a supported GPU will want to turn GPU acceleration on as will hugely improve processing times (which are on the slow side if only using CPU processing).
Speech Note is a 620MB download from Flathub (excluding any runtimes or platforms required) and takes up around ~2GB when installed – if you’re data or disk constrained, do keep those factors in mind.
• Get Speech Note on Flathub