Source: talky
Section: sound
Priority: optional
Maintainer: Talky Contributors <talky@example.com>
Build-Depends: debhelper-compat (= 13),
               dh-python,
               python3-all,
               python3-setuptools,
               python3-pip
Standards-Version: 4.6.0
Homepage: https://github.com/ChrisKalahiki/talky
Vcs-Browser: https://github.com/ChrisKalahiki/talky
Vcs-Git: https://github.com/ChrisKalahiki/talky.git
Rules-Requires-Root: no

Package: talky
Architecture: all
Depends: ${python3:Depends},
         ${misc:Depends},
         python3 (>= 3.10),
         python3-numpy,
         python3-pil,
         python3-yaml,
         python3-dotenv,
         xdotool | ydotool
Recommends: python3-torch,
            python3-sounddevice,
            python3-pystray,
            python3-pynput,
            python3-notify2,
            ydotool
Suggests: cuda-toolkit
Description: System-wide dictation for Linux using Whisper AI
 Talky is a system-wide dictation application for Linux, similar to
 WisprFlow AI on Windows/Mac. It uses OpenAI's Whisper model for
 accurate speech-to-text transcription and works across all applications.
 .
 Features:
  - Push-to-talk: Hold hotkey, speak, release
  - System-wide: Works in any application
  - Fast: <1.5s latency with CUDA GPU acceleration
  - Multi-language: Supports 99 languages
  - System tray with GUI
  - X11 and Wayland support
  - Privacy-focused: Local processing
