Voice Text Prog for Linux 1.0
Speak into your microphone and get Spanish text typed into any window.

Created by Leonel Prog · bitcoinprog.com
(Español: ver LÉEME.txt)

═══════════════════════════════════════════════════════════════
INSTALL
  ./instalar.sh
  Installs the missing system packages, creates its own Python
  environment and leaves a `voicetextprog` command ready, starting
  automatically when you log in.

  The FIRST time you dictate it downloads the Whisper model
  (~460 MB). After that it starts instantly.

HOW TO USE IT
  Ctrl+Alt+D ... start / finish, from ANY window
  While you speak a floating bar shows elapsed time and mic level.
  When you finish, the text types itself where your cursor was.

  In a terminal the text is flattened to a single line (so no line
  break runs a command) and pasted with Ctrl+Shift+V.

EVERYTHING STAYS ON YOUR MACHINE
  Whisper runs locally. Your audio never leaves your computer:
  not to Google, not to Apple, not to me.

NOTHING IS LOST
  ~/.local/share/voice-text-prog/dictados-YYYY-MM.txt

SETTINGS
  ~/.config/voice-text-prog/config.json
  model: tiny · base · small (default) · medium · large-v3
  Bigger = better Spanish, but slower and more RAM.

SELF-TEST (no microphone needed)
  voicetextprog --autoprueba audio.wav      (mono 16 kHz WAV)

REQUIREMENTS
  Linux with X11, Python 3.9+, a microphone.
  Works on Intel/AMD (x86_64) and ARM (aarch64).
  On Wayland the global shortcut may not work: bind a system
  shortcut to `voicetextprog --toggle` instead.
