Analysis updated 2026-05-18
Dictate a casual message with emoji into WhatsApp or Discord using a hotkey.
Dictate a formal email into Outlook with an appropriate greeting.
Speak inline editing commands like replace X with Y while dictating.
Correct a misheard word once and have the app apply that fix automatically going forward.
| mschr703/openwhisper | aburousan/typsteditor | deedy/glacial-valley | |
|---|---|---|---|
| Stars | 19 | 19 | 19 |
| Language | JavaScript | JavaScript | JavaScript |
| Setup difficulty | easy | moderate | easy |
| Complexity | 2/5 | 3/5 | 2/5 |
| Audience | general | researcher | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Node.js 18+ and an OpenAI account with billing set up for the Whisper API (~$0.006/minute).
Whisperlise (published here as OpenWhisper) is a desktop voice dictation app for Windows and macOS, built with Electron. You hold a keyboard shortcut, speak, and the app types the transcribed text directly into whatever application you are using at the moment, without you needing to switch windows or click a text field. It uses OpenAI's Whisper API to do the transcription, which costs roughly $0.006 per minute of recorded audio and requires an OpenAI account with billing set up. One of the main features is context awareness. The app detects which program is in the foreground and adjusts the tone of the transcribed text accordingly. If you are in a chat app like WhatsApp or Discord, it produces casual short text with emoji. If you are in an email client like Outlook, it produces more formal prose with appropriate greetings. If you are in a code editor, it preserves programming terms and formats comments correctly. If you are in a ChatGPT or Claude tab, it formats what you said as a clean prompt. While speaking, you can say inline editing commands such as "replace X with Y", "delete word", "new paragraph", or "format heading" and the app applies them to the current text on the fly. You can also end a dictation with a trigger phrase like "make professional" or "as bullet points" to reformat the entire thing before it is inserted. A self-learning correction dictionary lets you click on a word that Whisper got wrong in the history view to correct it, and the app then applies that correction automatically in future sessions. The overlay the app displays is a small transparent status capsule at the bottom of the screen that shows the current state (ready, listening, transcribing, typing, or done) without covering other windows or taking focus away from them. A settings dashboard lets you manage the correction dictionary, adjust the keyboard shortcut, pick a language, and view session history. Installation requires Node.js version 18 or higher. After cloning the repository and running "npm install", the app launches with "npm start". The default shortcut on Windows is hold Ctrl+Alt+Space to record and release to insert, on macOS it toggles on a single press.
A desktop voice dictation app that types transcribed speech into any app, adjusting tone and formatting based on which program is in focus.
Mainly JavaScript. The stack also includes JavaScript, Electron, OpenAI Whisper API.
Setup difficulty is rated easy, with roughly 30min to a first successful run.
Mainly general.
This repo across BitVibe Labs
Verify against the repo before relying on details.