WhisperWriter
An open-source dictation app that types speech into your active window using local Whisper models, CPU or NVIDIA processing, or OpenAI's API.
WhisperWriter turns microphone speech into text and types it into the window you're working in. It's for people who want voice input in their existing desktop apps, with a choice between transcription on their own computer and an external service. The Python app runs on Windows, macOS and Linux and uses the GPL-3.0 open-source license.
Local transcription is the default. It uses faster-whisper and can process audio on a CPU or an NVIDIA GPU. With a model stored locally, transcription keeps audio on your machine. You can choose Whisper model sizes to balance accuracy against speed, specify a language and supply a prompt to guide recognition.
For cloud transcription, the app sends recordings to OpenAI's API and requires an API key. It can also connect to a local API endpoint such as LocalAI, so the desktop dictation interface doesn't depend on running the model inside the app itself.
Recording suits both short phrases and longer dictation. A customizable keyboard shortcut controls listening, while continuous mode transcribes after a pause and resumes recording automatically. Other modes stop after silence, toggle recording with a key press, or record only while you hold the shortcut.
A small status window shows recording and transcription progress, and you can hide it. Text cleanup options include removing a final period, adding a trailing space and converting output to lowercase.
