Dictate text hands free into any application by holding a two-key combo and speaking.
Transcribe speech fully offline without sending audio to any cloud service.
Clean up dictated text's punctuation and sentence structure using a local Ollama model.
Set the tool to start automatically in the background whenever you log into Kali Linux.
| twinlights/wispr-clone | 0xallam/my-recipe | 0xhassaan/nn-from-scratch | |
|---|---|---|---|
| Stars | 0 | — | 0 |
| Language | Python | Python | Python |
| Last pushed | — | 2022-11-22 | — |
| Maintenance | — | Dormant | — |
| Setup difficulty | moderate | moderate | moderate |
| Complexity | 3/5 | 2/5 | 4/5 |
| Audience | developer | general | developer |
Figures from each repo's GitHub metadata at analysis time.
Requires Kali Linux with X11 (not Wayland) and downloading Whisper models while still online.
Wispr-clone is a dictation tool for Kali Linux that turns your voice into typed text, working completely offline and modeled after a commercial product called Wisprflow. You hold down two keys together, speak, and let go, and the tool types out what you said into whatever window has focus, without ever sending audio over the internet. The hold to talk gesture is deliberate: recording starts the moment both configured keys are pressed at the same time and stops the instant either one is released, rather than working like a toggle you press once to start and once to stop. The default combination uses the Windows key and AltGr specifically because those keys are almost never triggered by normal typing, which keeps the tool from firing by accident. Speech to text happens locally using a tool called faster-whisper, which understands both English and Dutch or Flemish and figures out which language you are speaking automatically. Whisper's own output already strips out filler words like "uh" and stutters. On top of that, an optional local step using Ollama, a tool for running small AI chat models on your own machine, can add proper punctuation, capitalization, and sentence structure, though it never translates the text and the tool still works fine without it if it is turned off or unreachable. The interface is a small oval widget built with Tkinter, using only a few megabytes of memory rather than a full browser based interface. The project includes an install script for Kali Linux running the X11 display system, a script to download and cache all supported Whisper model sizes ahead of time so the tool never needs the internet again once set up, and an optional background service that starts the tool automatically when you log in. The README does not state a license for this project.
An offline, push-to-talk voice dictation tool for Kali Linux that types out speech using local Whisper transcription while a key combo is held down.
Mainly Python. The stack also includes Python, faster-whisper, Tkinter.
Setup difficulty is rated moderate, with roughly 30min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Don't trust strangers blindly. Verify against the repo.