Clarify is an open-source Windows app I built for dictation and text rewriting across desktop applications. Global shortcuts capture speech or selected text, then return the result to the app where the user was working.
It uses a bring-your-own-key model: users choose a supported AI provider and supply their own API key. A separate Local Whisper option runs speech recognition on the device without an API key. These are different processing modes, not a claim that every AI feature runs locally.
I own the product, desktop workflow, provider integrations and release engineering. The public repository includes the source, setup instructions and MIT license. The image above shows the real settings interface with a demonstration configuration and no credentials.